abdullah's picture
Add files using upload-large-folder tool
25b5648 verified
raw
history blame
59.9 kB
1
00:00:04,940 --> 00:00:07,280
ุจุณู… ุงู„ู„ู‡ ูˆุงู„ุญู…ุฏ ู„ู„ู‡ ูˆุงู„ุตู„ุงุฉ ูˆุงู„ุณู„ุงู… ุนู„ู‰ ุฑุณูˆู„ ุงู„ู„ู‡
2
00:00:07,280 --> 00:00:12,260
ุฃู‡ู„ุง ูˆุณู‡ู„ุง ุจูƒู… ููŠ ู…ุญุงุถุฑุชู†ุง ุงู„ู…ุณุชู…ุฑู‘ุฉ... ููŠ
3
00:00:12,260 --> 00:00:15,740
ู…ุญุงุถุฑุงุชู†ุง ุงู„ู…ุณุชู…ุฑู‘ุฉ ููŠ ู…ุณุงู‚ ุงู„ู€ data mining ูˆู…ุง ุฒู„ู†ุง
4
00:00:15,740 --> 00:00:20,980
ู†ุชูƒู„ู… ููŠ ุจุงุจ ุงู„ู€ classificationุŒ ูˆุจุงู„ุชุญุฏูŠุฏ ุณู†ุชูƒู„ู…
5
00:00:20,980 --> 00:00:24,660
ุงู„ูŠูˆู… ุนู„ู‰ decision tree induction. ูƒู†ุง ููŠ ุงู„ู…ุญุงุถุฑุงุช
6
00:00:24,660 --> 00:00:27,540
ุงู„ุณุงุจู‚ุฉุŒ ุฃูˆ ุงู„ู…ุญุงุถุฑุฉ ุงู„ุฃุฎูŠุฑุฉุŒ ุฃุถูู†ุง ุดุบู„ุฉ ุฌุฏูŠุฏุฉ. ูƒู†ุง
7
00:00:27,540 --> 00:00:31,040
ู†ุชูƒู„ู… ุนู„ู‰ ุงู„ู€ Naive Bayes. ูƒุงู†ุช ูุนู„ูŠู‹ุง
8
00:00:31,040 --> 00:00:34,320
ู‡ูŠ ูˆุงุญุฏุฉ ู…ู† ุงู„ู€ probabilistic approach ุงู„ู…ุณุชุฎุฏู…ุฉ
9
00:00:34,320 --> 00:00:36,800
ููŠ ุงู„ู€ machine learning ู…ู† ุฃุฌู„ ุงู„ู€ classification.
10
00:00:36,800 --> 00:00:41,680
ูˆู‚ู„ู†ุง ุฅู†ู‘ู‡ ูŠู„ุฒู…ู†ุง ุฃู† ู†ู‚ูˆู… ุจุญุณุงุจ ู„ู…ุฌู…ูˆุนุฉ ู…ู†
11
00:00:41,680 --> 00:00:44,300
probabilities. ุฃู†ุง ุนู†ุฏูŠ ุงู„ู€ instance ุงู„ู„ูŠ ุจุฏูŠ...
12
00:00:44,300 --> 00:00:46,600
ุงู„ู„ูŠ ู‡ูŠ ุงู„ู€ unseen instance ุงู„ู„ูŠ ุฃู†ุง ุจุฏูŠ ุฃุนู…ู„ ู„ู‡ุง
13
00:00:46,600 --> 00:00:50,440
classification ุจู†ุงุกู‹ ุนู„ู‰... ูˆุจุงู„ุชุงู„ูŠ ุงู„ู€ class
14
00:00:50,440 --> 00:00:54,380
ุงู„ุทุจุนุฉ ู„ู„ู€ instance ู‡ุฐู‡ ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ุชุณุงูˆูŠ
15
00:00:54,380 --> 00:00:58,370
maximum probability ู„ู„ู€ probabilities of the class.
16
00:00:58,370 --> 00:01:05,470
ููŠ ุงุญุชู…ุงู„ูŠุฉ ุฃู† ุชูƒูˆู† ุงู„ู€ instance ู‡ุฐู‡ ู…ุน ุงู„ู€ class
17
00:01:05,470 --> 00:01:10,110
ุงู„ู…ุนูŠู†. ูˆู„ู…ุง ุฑุฌุนู†ุง ุจุงู„ุชูุตูŠู„ุŒ ู‚ู„ุช ุฃู†ุง ูุนู„ูŠู‹ุง ุจุญุงุฌุฉ
18
00:01:10,110 --> 00:01:14,070
ุฃู† ุขุฎุฐ ู‡ุฐู‡ ุงู„ู€ data setุŒ ูˆุฃุฐู‡ุจ ู„ุฃู†ุดุฆ ุงู„ุฌุฏูˆู„ ู‡ุฐุง ุจุญูŠุซ ุฃู†ู†ูŠ
19
00:01:14,070 --> 00:01:18,550
ุฃุฐู‡ุจุŒ ุญุณุจุช ุงู„ู€ probability ู„ูƒู„ element ุฃูˆ ู„ูƒู„
20
00:01:18,550 --> 00:01:21,770
classes ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ. ูˆู…ู† ุซู… ุงู†ุชู‚ู„ู†ุง ููŠ ุงู„ุฎุทูˆุฉ
21
00:01:21,770 --> 00:01:24,910
ุงู„ุชูŠ ุจุนุฏู‡ุงุŒ ุฃุฎุฐุช ุงู„ู€ attributes ุงู„ุชูŠ ู…ู† ุงู„ู…ูุชุฑุถ
22
00:01:24,910 --> 00:01:27,370
ุงู„ุชูŠ ู…ู† ุงู„ู…ูุชุฑุถ ุฃู† ูŠูƒูˆู† ู„ุฏูŠู‡ุง nominal attributesุŒ ุฃุฎุฐุช
23
00:01:27,370 --> 00:01:30,950
distinct valuesุŒ ูˆุนู…ู„ุช ุญุณุงุจ ู„ูƒู„ ูˆุงุญุฏุฉ ู…ู†ู‡ู…ุŒ ูˆุงู†ุชุจู‡ูˆุง
24
00:01:30,950 --> 00:01:35,520
ุฏุงุฆู…ู‹ุงุŒ ูˆุงู†ุชุจู‡ูˆุง ุฏุงุฆู…ู‹ุง ุฃู†ู†ูŠ ูุนู„ูŠู‹ุง ู‡ู†ุง ู‚ุงุนุฏ ุจุฃุดุชุบู„ ุนู„ู‰
25
00:01:35,520 --> 00:01:38,840
ุฃู† ุงู„ู€ probability ู†ูุณู‡ุงุŒ ูŠุนู†ูŠ ุงู„ุขู† ุนุฏุฏ ุงู„ู€ yes ููŠ ุงู„ู€
26
00:01:38,840 --> 00:01:42,880
dataุŒ ุงู„ู„ูŠ ุนู†ุฏูŠ ู‡ู†ุง 4 ุนู„ู‰ 10. ูˆู…ู† ุซู… ู…ุน ูƒู„ route
27
00:01:42,880 --> 00:01:48,400
ุฃูˆ ูƒู„ attribute ุชุญุช ุงู„ู€ yes ุณูŠูƒูˆู† 4ุŒ ูˆูƒู„ ู…ุฌู…ูˆุน
28
00:01:48,400 --> 00:01:51,920
ุงู„ุนู†ุงุตุฑ ุชุญุช ูƒู„ no ุณูŠูƒูˆู† 6ุŒ ูˆู‡ูƒุฐุง. ูˆู‡ุฐุง ู…ูุชุงุญ
29
00:01:51,920 --> 00:01:56,510
ุงู„ู†ุฌุงุญุŒ ู„ู„ุนู†ุงุตุฑ ุงู„ู…ูˆุฌูˆุฏุฉ. ูˆู„ู…ุง ุฌุฆู†ุง ู†ูุตู†ู‘ู ุงู„ู€
30
00:01:56,510 --> 00:01:59,230
sunny ูˆุงู„ู€ mild ูˆุงู„ู€ highุŒ ู‚ู„ู†ุง ุญุณุจุช ุงู„ู€
31
00:01:59,230 --> 00:02:01,670
probability ู„ู„ู€ yes ุงู„ุชูŠ ูƒุงู†ุช 4 ุนู„ู‰ 10 ููŠ ุงู„ู€
32
00:02:01,670 --> 00:02:04,690
probability ู„ู„ู€ sunny ุนู„ู‰ ุงู„ู€ yesุŒ ูˆู‚ู„ู†ุง ู‡ุฐุง ุงู„ุฌุฏูˆู„
33
00:02:04,690 --> 00:02:07,810
ุงู„ุฃุณุงุณ ููŠ ุงู„ู…ูˆุถูˆุนุŒ ู‡ูŠ sunny ูˆ yesุŒ ู‡ูŠู‡ุง 4 ุนู„ู‰ 10
34
00:02:07,810 --> 00:02:12,530
ู…ุถุฑูˆุจุฉ ููŠ ุงู„ุนู†ุตุฑ ุงู„ุซุงู†ูŠุŒ ูƒุงู†ุช mildุŒ ุงู„ู€ probability
35
00:02:12,530 --> 00:02:17,190
ุชุจุนุช ุงู„ู€ mildุŒ ูŠุนู†ูŠ ุจูŠู† ู‚ูˆุณูŠู† ู‡ูŠ ุงู„ู€ yes ู‡ุฐู‡ ู…ุถุฑูˆุจุฉ
36
00:02:17,190 --> 00:02:22,410
ููŠ ู‡ุฐู‡ ููŠ ุงู„ู€ mild ููŠ ุงู„ู€ highุŒ ูˆู‡ุฐู‡ ุงู„ุนู†ุงุตุฑ ูƒุงู†ุช
37
00:02:22,410 --> 00:02:25,050
ุชู…ุซู‘ู„ ุงู„ู€ probability. ูุฃู†ุง ุญุณุจุช ุงู„ู€ probability ู„ู„ู€
38
00:02:25,050 --> 00:02:27,850
different classes ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠุŒ ูˆุฃุฎุฐุช ุงู„ู€
39
00:02:27,850 --> 00:02:31,810
maximum probabilityุŒ ุนู„ู‰ ุฃู† ู‡ุฐู‡ ู‡ูŠ ุงู„ุฃูƒุซุฑ ุงุญุชู…ุงู„ูŠุฉ
40
00:02:31,810 --> 00:02:36,090
ููŠ ู…ูˆุถูˆุน ุฃู† ู‡ุฐุง ุงู„ุนู†ุตุฑ ุฃูˆ ู‡ุฐู‡ ุงู„ู€ instance ุชู†ุชู…ูŠ ู„ู„ู€
41
00:02:36,090 --> 00:02:40,850
class ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ุง. ุงู„ุขู† ุงู„ูŠูˆู… ุฅู† ุดุงุก ุงู„ู„ู‡
42
00:02:40,850 --> 00:02:45,620
ุณู†ุชู†ุชู‚ู„ ุฅู„ู‰ ุงู„ู…ูˆุถูˆุน ุงู„ุฌุฏูŠุฏุŒ ูˆู‡ูˆ ู…ูˆุถูˆุน ุงู„ู€
43
00:02:45,620 --> 00:02:49,500
decision tree ููŠ ุงู„ุชุนุงู…ู„ ุฃูˆ ูƒู€ different classifier.
44
00:02:49,500 --> 00:02:53,300
ุงู„ู€ decision tree ู‡ูŠ ูˆุงุญุฏุฉ ู…ู† ุงู„ู€ classifiers ุงู„ู…ู‡ู…ุฉ
45
00:02:53,300 --> 00:02:57,200
ุฌุฏุงุŒ ุงู„ู…ุณุชุฎุฏู…ุฉ ููŠ ู…ูˆุถูˆุน ุงู„ู€ classificationุŒ ูˆุฃู‡ู…ูŠุชู‡ุง
46
00:02:57,200 --> 00:03:00,840
ุฃู†ู‘ูŠ ุฃู…ูƒู†ู†ูŠ ุฃู† ุฃุฑุณู… ุงู„ุดุฌุฑุฉุŒ ูˆุจุงู„ุชุงู„ูŠ ูŠุตูŠุฑ ุชูุณูŠุฑ
47
00:03:00,840 --> 00:03:04,000
ุงู„ู€ model ุงู„ู…ูˆุฌูˆุฏ ุนู†ุฏูŠุŒ ุฃูˆ ูู‡ู… ุงู„ู€ model ุงู„ุฐูŠ ุนู†ุฏูŠ
48
00:03:04,000 --> 00:03:07,210
ุฃูƒุซุฑ ู…ู† ุบูŠุฑู‡. ุนู„ู‰ ุณุจูŠู„ ุงู„ู…ุซุงู„ุŒ ู‚ู„ู†ุง ููŠ ุงู„ู€
49
00:03:07,210 --> 00:03:10,610
classifier ุงู„ู…ุงุถูŠุŒ ูˆู‡ูˆ naive BayesุŒ ุฃู†ู†ูŠ ูุนู„ูŠู‹ุง
50
00:03:10,610 --> 00:03:13,970
ุนู†ุฏ ุงู„ู€ classifier ู‡ุฐุงุŒ ู…ู‡ู… ุฃูˆ ุฌูŠุฏุŒ ู„ุฃู†ู‡ ุฃู†ุง ุฃู‚ุฏุฑ
51
00:03:13,970 --> 00:03:17,090
ุฃูุณุฑ ู„ู…ุงุฐุง ุงู„ู†ุชูŠุฌุฉ ุทู„ุนุช ู…ุนูŠ ู‡ูƒุฐุงุŒ ุจู†ุงุกู‹ ุนู„ู‰
52
00:03:17,090 --> 00:03:21,310
ุงู„ุงุญุชู…ุงู„ุงุช ุงู„ู…ูˆุฌูˆุฏุฉ. ููŠ ุงู„ู€ decision tree ูƒุฐู„ูƒุŒ ููŠ
53
00:03:21,310 --> 00:03:24,590
decision tree ู‡ูŠ ุนุจุงุฑุฉ ุนู† ุฃู†ู†ูŠ ูุนู„ูŠู‹ุง ุฃูุจู†ูŠ
54
00:03:24,590 --> 00:03:26,930
decision tree. ุฃูŠู‡ุง ุงู„ุณุงุฏุฉุŒ ู„ู…ุง ุฃู†ุง ุจุฃุชูƒู„ู… ุนู„ู‰
55
00:03:26,930 --> 00:03:31,210
decision treeุŒ ุชุฐูƒุฑูˆุงุŒ ุฏุนูˆู†ุง ู†ุชุฐูƒุฑ ุจุณุฑุนุฉ ุงู„ู€
56
00:03:31,210 --> 00:03:34,930
binary search tree. ู„ุง ุฃุฑูŠุฏ ุฃูƒุซุฑ ู…ู† ุฐู„ูƒ. ุงู„ู€ binary
57
00:03:34,930 --> 00:03:38,490
search tree ูƒุงู†ุช ุนู†ุงุตุฑู‡ุง ุฃู†ู‘ ูƒู„ node ุนู„ู‰
58
00:03:38,490 --> 00:03:43,580
ุงู„ุฃูƒุซุฑ ู„ุฏูŠู‡ุง two childุŒ ุตุญูŠุญุŸ ู‡ุฐู‡ ู‡ูŠ ุงู„ู€ binary tree.
59
00:03:43,580 --> 00:03:47,140
ูˆูƒุงู† ููŠู‡ rule ูŠุญูƒู…ู‡ุงุŒ ุงู„ู€ rule ุฃู†ู†ูŠ ููŠ ุงู„ู€ binary
60
00:03:47,140 --> 00:03:51,200
search treeุŒ ุฃู†ู‘ ูƒู„ ุงู„ู‚ูŠู… ุงู„ุชูŠ ุนู„ู‰ ุงู„ูŠู…ูŠู† ู‡ู†ุง ุณุชูƒูˆู†
61
00:03:51,200 --> 00:03:55,570
ุฃูƒุจุฑ ู…ู† ุงู„ู€ elementุŒ ูˆูƒู„ ุงู„ู‚ูŠู… ุงู„ุชูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ุง ู‡ู†ุง
62
00:03:55,570 --> 00:03:59,330
ุณุชูƒูˆู† ุฃุตุบุฑ. ุทูŠุจุŒ ุงู„ู‚ูŠู… ุงู„ู…ุชุณุงูˆูŠุฉุŸ ู…ุงู„ู‡ุง ูˆุฌูˆุฏุŒ ุงู„ู…ูƒุฑุฑุฉ
63
00:03:59,330 --> 00:04:02,830
ู…ุงู„ู‡ุง ูˆุฌูˆุฏ. ูˆุจุงู„ุชุงู„ูŠ ุงู„ู€ element ู…ุน ูƒู„ nodeุŒ ุงู„ู€ node
64
00:04:02,830 --> 00:04:05,690
ุงู„ุชูŠ ุนู†ุฏู‡ุง ุงู„ู‚ูŠู… ุงู„ู…ูˆุฌูˆุฏุฉ ู‡ู†ุง ุณุชูƒูˆู† ุฃุตุบุฑ ู…ู† ุงู„ู‚ูŠู…
65
00:04:05,690 --> 00:04:08,810
ุงู„ุชูŠ ู…ูˆุฌูˆุฏุฉ ู‡ู†ุงุŒ ูˆู‡ุฐู‡ ุทุจูŠุนุชู‡ุง ุณุชูƒูˆู† ุฃุตุบุฑ ู…ู† ุงู„ู‚ูŠู…
66
00:04:08,810 --> 00:04:12,330
ุงู„ุชูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ุง. ุจู…ุนู†ู‰ ุขุฎุฑุŒ ุฃู†ู‘ ุงู„ู€ structure ุชุจุน ุงู„ู€
67
00:04:12,330 --> 00:04:15,250
decision treeุŒ ุฃู†ุง ุจุงู„ูุนู„ ุฃุนุฑูู‡ุงุŒ ู‡ูŠ ุนุจุงุฑุฉ ุนู†
68
00:04:15,250 --> 00:04:18,270
ู…ุฌู…ูˆุนุฉ ู…ู† ุงู„ู€ nodes ูˆุงู„ู€ connection ุฃูˆ ุงู„ู€ connected
69
00:04:18,270 --> 00:04:22,790
ุฃูˆ ุงู„ู€ graph with no circuitุŒ ุฒูŠ ู…ุง ูƒู†ุง ู†ุณู…ูŠู‡ุง ููŠ ุงู„ู€
70
00:04:22,790 --> 00:04:25,670
discrete mathematicsุŒ ุฃุดุจู‡ ุจุงู„ู€ flow chartุŒ ุฒูŠ ู…ุง
71
00:04:25,670 --> 00:04:28,750
ู‚ู„ู†ุง ุณุงุจู‚ู‹ุง. ููŠ ุนู†ุฏูŠ ุงู„ู€ internal nodeุŒ ูˆุงู„ู„ูŠ ุฃู†ุง
72
00:04:28,750 --> 00:04:32,150
ูุนู„ูŠู‹ุง ู‡ูŠ ุงู„ู€ value ุชุจุนู‡ุง... ุณุชู…ุซู„ ุงู„ู€ value ุชุจุน ุงู„ู€
73
00:04:32,150 --> 00:04:35,490
attribute ุงู„ุชูŠ ุณุฃุญู…ู„ ุนู„ูŠู‡ุง ุงู„ูุญุตุŒ ุณุฃุณุฃู„: ูƒู… ุงู„ู€ GPA
74
00:04:35,490 --> 00:04:40,550
ุฃูƒุจุฑ ุฃูˆ ุชุณุงูˆูŠ ูƒุฐุงุŒ ุฃุฐู‡ุจ ูŠู…ูŠู†ู‹ุงุŒ ุฃู‚ู„ ุฃูˆ falseุŒ ุฃุฐู‡ุจ ูŠุณุงุฑู‹ุง.
75
00:04:40,550 --> 00:04:45,090
ูˆู‡ูƒุฐุง. ูู‡ุฐู‡ ุงู„ู€ internal nodeุŒ ุงู„ุชูŠ ู‡ูŠ ุนุงุฏุฉ non-leafุŒ
76
00:04:45,950 --> 00:04:49,790
ุชุญุฏุฏ ุงู„ู€ test ุชุจุน ุงู„ู€ attributeุŒ ุงู„ู€ branch ูŠู…ุซู„ ุงู„ู€
77
00:04:49,790 --> 00:04:53,210
outcomeุŒ ูˆุตูˆู„ู‹ุง ู„ู„ู€ leaf ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠุŒ ูˆุงู„ู€ leaf
78
00:04:53,210 --> 00:04:58,070
node ุชู…ุซู‘ู„ ุงู„ู€ classุŒ ูˆุทุจุนู‹ุง ู„ุงุฒู… ูƒู„ ุดุฌุฑุฉ ูŠูƒูˆู† ู„ู‡ุง
79
00:04:58,070 --> 00:05:03,690
root node. ุชุนุงู„ูˆุง ู†ุดูˆู ุงู„ู€ data set ุงู„ุจุณูŠุทุฉ ุงู„ุชูŠ
80
00:05:03,690 --> 00:05:08,870
ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู†ุง. ู‡ุฐู‡ ุงู„ู€ data set ู…ูƒูˆู‘ู†ุฉ ู…ู† 14 rowุŒ ุงู„ู€
81
00:05:08,870 --> 00:05:12,490
Age ูˆุงู„ู€ income ูˆุงู„ู€ student ูˆุงู„ู€ credit rating ูˆุงู„ู€
82
00:05:12,490 --> 00:05:15,710
class ุชุจุนูŠ. ูˆุทุจุนู‹ุงุŒ ุฃูŠู‡ุง ุงู„ุณุงุฏุฉุŒ ู„ู…ุง ู†ุญู† ู†ุฐู‡ุจ
83
00:05:15,710 --> 00:05:20,610
ุจุงุชุฌุงู‡ ุงู„ู€ binary classุŒ ูŠุนู†ูŠ two classesุŒ ุงู„ุฃู…ุฑ ุงู„ุฐูŠ
84
00:05:20,610 --> 00:05:23,690
ุฃุณู‡ู„ุŒ ู„ุฃุฌู„ ุงุณุชูŠุนุงุจ ู…ุงุฐุง ูŠุญุฏุซุŒ ู„ุฃู†ู‡ ู„ู…ุง ุชุตุจุญ ุนู†ุฏูŠ
85
00:05:23,690 --> 00:05:26,430
ุซู„ุงุซุฉุŒ ุณุชุชุดุนุจ ุงู„ุฃู…ูˆุฑ ู‚ู„ูŠู„ู‹ุง ููŠ ุงู„ุญุณุงุจุŒ ู„ูƒู†ู‘ู‡ุง
86
00:05:26,430 --> 00:05:32,030
ุนุจุงุฑุฉ ุนู† ุชูƒุฑุงุฑ ู„ู…ุง ุณุจู‚. ุงู„ุขู† ุฃู†ุง... ุนู†ุฏูŠ ู…ุฌู…ูˆุนุฉ
87
00:05:32,030 --> 00:05:35,530
ู…ู† ุงู„ุทู„ุงุจุŒ ุฃูˆ ุจูŠุงู†ุงุช ู…ุฌู…ูˆุนุฉ ู…ู† ุงู„ู†ุงุณ ุงู„ุฐูŠู† ุงุดุชุฑูˆุง
88
00:05:35,530 --> 00:05:39,190
ุญุงุณุจุงุช. ูˆุงู„ู€ data set ูƒุงู†ุช ู‚ุฏูŠู…ุฉุŒ ููŠ ุงู„ู€ 2000ุŒ ูˆูƒุงู†
89
00:05:39,190 --> 00:05:42,630
ูŠุณุฃู„: ู‡ู„ ู‡ุฐุง ุงู„ุดุฎุต ู…ุคู‡ู‘ู„ุŒ ุฃูˆ ู…ู…ูƒู†ุŒ ู…ุน ุงุญุชู…ุงู„ ุฃู† ูŠุดุชุฑูŠ
90
00:05:42,630 --> 00:05:47,430
ุฌู‡ุงุฒู‹ุง ุฃู… ู„ุงุŒ ุจู†ุงุกู‹ ุนู„ู‰ ุญุงู„ุชู‡. ุงู„ู€ ageุŒ ู„ุงุญุธูˆุง ุงู„ู€ ageุŒ ุฃู†ุง
91
00:05:47,430 --> 00:05:52,190
ุฃุชูƒู„ู… ุนู„ู‰ discrete ุฃูˆ categorial data. ุงู„ู€ incomeุŒ
92
00:05:52,190 --> 00:05:56,210
high ูˆ low ูˆ medium. ุทุงู„ุจ ุฃู… ุบูŠุฑ ุทุงู„ุจุŒ yes ุฃูˆ no. ูˆุงู„ู€
93
00:05:56,210 --> 00:06:00,750
credit ratingุŒ ุนุงุฏูŠ ุฃูˆ ู…ุนุชุฏู„ ุฃูˆ excellent.
94
00:06:00,750 --> 00:06:03,830
ุจุงู„ู†ุณุจุฉ ู„ู„ู…ุชูˆุณุท ุงู„ุฑุงุชุจ ุชุจุนู‡ุŒ ูˆููŠ ุงู„ุขุฎุฑ ุงู„ู€ class
95
00:06:03,830 --> 00:06:07,330
ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ู…. ุงู„ุขู†ุŒ ู„ู…ุง ุฃู†ุง ุจุฏูŠ ุฃุจู†ูŠ treeุŒ ุงู„ู€
96
00:06:07,330 --> 00:06:11,730
tree ุณุชุฃุฎุฐ ุจุงู„ุดูƒู„ ู‡ุฐุง. ุฏุนูˆู†ุงุŒ ู„ุฃุฌู„ ุฃู† ู†ุฃุฎุฐ ุนู„ู‰
97
00:06:11,730 --> 00:06:16,110
ุงู„ุณุฑูŠุนุŒ ู†ุฃุฎุฐ ุงู„ู€ role ุฃูˆู„ู‹ุงุŒ ู„ุฃุฌู„ ุฃู† ู†ุชุฐูƒุฑุŒ ูˆู†ุดูˆู ูƒูŠู ุจุฏูŠ
98
00:06:16,110 --> 00:06:19,350
ุฃุชู†ุงูˆู„ ุงู„ู€ treeุŒ ุฃูˆ ูƒูŠู ุชุนู…ู„ ูƒู€ classifier. yesุŒ
99
00:06:19,350 --> 00:06:24,090
highุŒ noุŒ fair.
100
00:06:26,540 --> 00:06:32,680
ุงู„ู€ target ุชุจุนูŠุŸ ู„ุงุŒ ู‡ุฐุง ุฃูˆู„ row. ุจุณ ุฃู†ุงุŒ ู„ุฃุฌู„ ุฃู† ุฃุบูŠู‘ุฑุŒ
101
00:06:32,680 --> 00:06:36,700
ุจุฏูŠ ุฃุญุท ู‡ู†ุง yesุŒ ู„ุฃุฌู„ ุฃู† ุชุตุจุญ ู‡ุฐู‡ ุงู„ู€ data ุฃุดุจู‡ ุจุงู„ู€
102
00:06:36,700 --> 00:06:41,660
unseenุŒ ูˆุฃุดูˆู ุจุงู„ู€ classification ุชุจุนู‡ุง ูƒูŠู ุจุฏู‡ุง
103
00:06:41,660 --> 00:06:47,620
ุชูƒูˆู†. ุงู„ุขู†ุŒ ุฒูŠ ู…ุง ู‚ู„ู†ุงุŒ ู„ุฏูŠู†ุง ุงู„ู€ ageุŒ ุงู„ู€ incomeุŒ
104
00:06:47,620 --> 00:06:50,860
studentุŒ
105
00:06:50,860 --> 00:06:54,920
ูˆููŠ ุงู„ุขุฎุฑ ุงู„ุชูŠ ู‡ูŠ ุงู„ู€ credit
106
00:07:02,220 --> 00:07:06,100
rate. ุญุงุฌุฉ ุนู„ู‰ decision tree. ุงู„ู€ decision tree ุฃูˆ ุงู„ู€
107
00:07:06,100 --> 00:07:09,400
modelุŒ ู„ู…ุง ุชู… ุจู†ุงุคู‡ุŒ ุฌุงุก ูŠู‚ูˆู„: ุฃู‡ู… element ููŠ
108
00:07:09,400 --> 00:07:13,660
ุงู„ู‚ุฑุงุฑ ุนู†ุฏูŠุŒ ุงู„ู€ age. ูˆุณู†ุชุนุฑู‘ู ูƒู…ุงู† ู„ุญุธุงุชุŒ ุฅู† ุดุงุก ุงู„ู„ู‡
109
00:07:13,660 --> 00:07:17,240
ุชุนุงู„ู‰ุŒ ูƒูŠู ู†ุญู† ุงุฎุชุฑู†ุง ุงู„ู€ age. ู„ู…ุงุฐุง ู„ู… ุชูƒู†
110
00:07:17,240 --> 00:07:20,060
studentุŒ ุฃูˆ ู„ู€ credit ratingุŒ ุงู„ุชูŠ ู‡ูŠ ุงู„ู€ attributes
111
00:07:20,060 --> 00:07:23,260
ุงู„ุฃุฎุฑู‰ุŸ ูˆู„ุงุญุธูˆุง ุฃู†ู‘ ุนู†ุฏูŠุŒ ุจุงู„ูƒุงู…ู„ุŒ ุนู†ุฏูŠ
112
00:07:23,260 --> 00:07:27,950
attribute ุบุงุฆุจุŒ ูˆู‡ูˆ ู…ูˆุถูˆุน ุงู„ู€ income ููŠ ุงู„ู€
113
00:07:27,950 --> 00:07:30,530
decision tree ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ. ูŠุนู†ูŠ ู‡ูˆ ูƒุฃู†ู‡ ูŠู‚ูˆู„
114
00:07:30,530 --> 00:07:35,270
ุงู„ู€ income ู‡ู†ุง ู„ูŠุณ ู„ู‡ ุชุฃุซูŠุฑ ูƒุซูŠุฑ ุนู„ู‰ ุงู„ู€ decision
115
00:07:35,270 --> 00:07:39,010
ุฃูˆ ุนู„ู‰ ุงู„ู‚ุฑุงุฑ ุงู„ู…ูˆุฌูˆุฏ ุนู†ุฏูŠ. ุทูŠุจุŒ ูƒูŠู ู‡ุฐุง ุงู„ูƒู„ุงู…
116
00:07:39,010 --> 00:07:44,590
ุตุงุฑุŸ ุณู†ุฑุงู‡ ู‚ู„ูŠู„ู‹ุงุŒ ู„ู…ุงุฐุงุŸ ููŠ ุฏู„ุงู„ุฉ ุฃุฎุฑู‰ ู‡ู†ุงุŒ ูŠุฃุชูŠ
117
00:07:44,590 --> 00:07:47,350
ูŠู‚ูˆู„ ุฅู†ู‘ ุงู„ู€ income ุงู„ุฃู‚ู„ ุชุฃุซูŠุฑู‹ุงุŒ ุฃูˆ ู„ูŠุณ ู„ู‡ ุชุฃุซูŠุฑุŒ ูŠุนู†ูŠ
118
00:07:47,350 --> 00:07:50,130
ุจุฅู…ูƒุงู†ูŠ ุฃู† ุฃุดูŠู„ู‡ุŒ ุฃูˆ ุฃุณุชุบู†ูŠ ุนู†ู‡. ุจุฌู…ุน ุงู„ุจูŠุงู†ุงุช ุงู„ุฃุฎุฑู‰
119
00:07:50,130 --> 00:07:55,510
ุณุชูƒูˆู† ุฃุณู‡ู„. ุงู„ุขู†ุŒ ุงู„ู€ age: youth, middle ageุŒ ูˆ senior.
120
00:07:55,510 --> 00:07:59,210
ููŠ ุนู†ุฏูŠ ุชูุฑุนุงุช ุฃุฎุฑู‰ุŸ ู„ุงุŒ ู‡ุคู„ุงุก ุงู„ู€ three discrete
121
00:07:59,210 --> 00:08:02,850
values ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ุจุงู„ู€ ageุŒ ุชู…ุงู…ุŸ ุญุณุจ ุงู„ู€ rule
122
00:08:02,850 --> 00:08:09,970
ุนู†ุฏู†ุง ู‡ู†ุงุŒ ู‡ูŠ ุงู„ู€ age. ุฅุฐุง ุฃู†ุง ุณุฃู„ุชู‡ ageุŒ ูŠุนู†ูŠ ุฎู„ุงุต ูƒู„
123
00:08:09,970 --> 00:08:14,010
branch ุงู„ุณุงุจู‚ ู‡ุฐู‡ ู„ูŠุณ ู„ูŠ ุฏุฎู„ ุจู‡ุง. ุฅุฐุง ูƒุงู† ู‡ูˆ
124
00:08:14,010 --> 00:08:19,770
studentุŒ ุบุงู„ุจู‹ุง ุณูŠุดุชุฑูŠุŒ ูˆู„ู† ุฃุจุญุซ ุนู† ุงู„ู…ุฒูŠุฏุŒ ู„ู†
125
00:08:19,770 --> 00:08:22,210
ุชุญุชุงุฌู†ูŠ. ุชุนุงู„ูˆุงุŒ ุทู„ุน ู…ุนูŠ ู‡ู†ุง.
126
00:08:26,000 --> 00:08:28,800
ูˆู‡ุฐุง ุงู„ุทุงู„ุจ ุณูŠูƒูˆู† ูŠุดุชุฑูŠ ูƒู…ุจูŠูˆุชุฑุŒ ู„ู…ุงุฐุงุŸ ู„ุฃู†ู‘ ุฅุฐุง
127
00:08:28,800 --> 00:08:31,800
ูƒุงู† ู‡ูˆ ููŠ ุงู„ู€ middle ageุŒ ุฃูˆ ููŠ ุงู„ู€ youthุŒ ุตุบูŠุฑ ุฃูˆ ุดุงุจ
128
00:08:31,800 --> 00:08:35,300
ูŠุงูุน ูˆุทุงู„ุจ ููŠ ู†ูุณ ุงู„ูˆู‚ุชุŒ ูŠุนู†ูŠ ุทุงู„ุจ ุฌุงู…ุนุฉุŒ ูุบุงู„ุจู‹ุง
129
00:08:35,300 --> 00:08:38,820
ู‡ุฐุง ุณูŠุญุชุงุฌ ูƒู…ุจูŠูˆุชุฑุŒ ูˆู…ู† ุซู… ุณูŠุฐู‡ุจ ู„ูŠุดุชุฑูŠู‡. ู„ูˆ ุฃู†ุง ุจุฏูŠ
130
00:08:38,820 --> 00:08:42,000
ุฃุนูˆุฏ ุฅู„ู‰ ุงู„ู€ data set ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ู‡ู†ุงุŒ youth ูˆ
131
00:08:42,000 --> 00:08:49,200
student ูˆ fairุŒ yes. youthุŒ studentุŒ ูˆ fairุŒ
132
00:08:51,960 --> 00:08:54,880
ู„ุฃุฌู„ ุฃู† ุชู„ุงุญุธูˆุง ุฃู†ู‘ ู‡ุฐูŠู† ู‡ู…ุง ุงู„ู„ุฐุงู† ูƒุงู†ุง ุฃูƒุซุฑ
133
00:08:54,880 --> 00:09:00,340
ุชุฃุซูŠุฑู‹ุง ููŠ ุญุงู„ุฉ ุงู„ู€ elements ุงู„ู…ูˆุฌูˆุฏุฉ. ูˆู‡ูƒุฐุงุŒ ู„ูˆ
134
00:09:00,340 --> 00:09:05,040
ูƒุงู† still age ููŠ ุงู„ู€ middle ageุŒ ู…ุจุงุดุฑุฉ ุณูŠูƒูˆู† ูŠุดุชุฑูŠ
135
00:09:05,040 --> 00:09:08,040
ุงู„ู€ attributeุŒ ู„ูˆ ูƒุงู† seniorุŒ
136
00:09:10,850 --> 00:09:14,250
ูˆุงู„ู€ income rate ุงู„ุฐูŠ ุนู†ุฏู‡ fairุŒ ุบุงู„ุจู‹ุง ู„ู† ูŠุดุชุฑูŠู‡.
137
00:09:14,250 --> 00:09:17,210
ูˆู‡ูƒุฐุง ุชุตุจุญ ู…ูˆุถูˆุน ุงู„ู€ decision ุฃูˆ ู…ูˆุถูˆุน ุงู„ู€
138
00:09:17,210 --> 00:09:19,950
classification. ูŠุนู†ูŠ ุงู„ู€ leaves ุงู„ุชูŠ ุนู†ุฏูŠ ููŠ ุงู„ู€ node
139
00:09:19,950 --> 00:09:23,850
ุฃูˆ ุนููˆุงู‹ ููŠ ุงู„ู€... ููŠ ุงู„ู€ treesุŒ ุงู„ู€ leaf nodes ุชู…ุซู‘ู„
140
00:09:23,850 --> 00:09:27,930
ุงู„ู€ classes ุงู„ุชูŠ ุฃู†ุง ุฃุจุญุซ ุนู†ู‡ุง. ูˆุทุจุนู‹ุง ุนู…ู‚
141
00:09:27,930 --> 00:09:33,210
ุงู„ุดุฌุฑุฉ ูˆุญุฌู…ู‡ุง ู…ุฑุชุจุท ุจุนุฏุฏ ุงู„ู€ attributes ูˆุญุฌู… ุงู„ู€
142
00:09:33,210 --> 00:09:35,970
data set ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ.
143
00:09:38,250 --> 00:09:42,710
ุงู„ุขู†ุŒ ุงู„ู€ algorithm ุงู„ู…ุณุชุฎุฏู… ู…ุน ุงู„ู€ decision treeุŒ
144
00:09:42,710 --> 00:09:50,490
ูˆู‡ูˆ ุงู„ู€ basic algorithmุŒ ู†ุณู…ูŠู‡ C4.5. ูˆู‡ุฐุง ูŠุนู…ู„
145
00:09:50,490 --> 00:09:54,150
ููŠ ู…ุจุฏุฃ ุงู„ู€ top-down recursive divide and conquer.
146
00:09:54,150 --> 00:09:58,730
ุงู„ุขู†ุŒ ุงู„ู†ุงุณ ุงู„ุฐูŠู† ุฃุฎุฐูˆุง ุฎูˆุงุฑุฒู…ูŠุงุชุŒ ุญุชู…ู‹ุง ู…ุฑู‘ ุนู„ูŠู‡ู… ู…ุตุทู„ุญ
147
00:09:58,730 --> 00:10:03,170
divide and conquer. ุงู„ููƒุฑุฉ ููŠ ุงู„ู€ algorithm ู‡ุฐุงุŒ ุฃู†ู‘
148
00:10:03,170 --> 00:10:07,810
ุงู„ู…ุดูƒู„ุฉ ุงู„ูƒุจูŠุฑุฉ ู†ุฌุฒู‘ุฆู‡ุงุŒ ูˆู†ุณุชุทูŠุน ุงู„ุณูŠุทุฑุฉ ุนู„ูŠู‡ุง. ูŠุนู†ูŠ
149
00:10:07,810 --> 00:10:11,630
ุจู†ุฌุฒุฆู‡ุงุŒ ู†ุญู„ ุฌุฒุก ุฌุฒุก ู…ู† ุงู„ู€ data set ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูƒ
150
00:10:11,630 --> 00:10:14,890
ู‡ู†ุงุŒ ูˆุณู†ุฑู‰ ู‚ู„ูŠู„ู‹ุง. ูˆู‡ุฐุง ุงู„ู…ุจุฏุฃ ู‡ูˆ ู…ุจุฏุฃ ุงู„ู€
151
00:10:14,890 --> 00:10:17,870
greedyุŒ ุทุจุนู‹ุง. ูˆู…ู† ุซู… ุงู„ู€ algorithm ู‡ุฐุง ุฃุฎุฐ ุงู„ู€ greedy
152
00:10:17,870 --> 00:10:21,450
algorithm. ู…ุงุฐุง ูŠุนู†ูŠ greedyุŸ ูŠุนู†ูŠ ุงู„ุทู…ุงุน. ููƒุฑุชู‡ุŒ ุจูƒู„
153
00:10:21,450 --> 00:10:24,710
ุจุณุงุทุฉุŒ ุฃู†ู†ูŠ ุฃู†ุธุฑ ุฅู„ู‰ best solution ููŠ ุงู„ู€ current
154
00:10:24,710 --> 00:10:29,210
stageุŒ ู„ุง ุฃู‡ุชู… ุนู„ู‰ ุงู„ู…ุฏู‰ ุงู„ุจุนูŠุฏุŒ ู…ุงุฐุง ุณูŠุญุฏุซ ุนู†ุฏูŠุŸ
155
00:10:30,520 --> 00:10:33,720
ุณุฃุจุฏุฃ ู…ุน ูƒู„ examplesุŒ ุณุขุฎุฐ ู…ุน ุงู„ู€ data ุงู„ู€ attributes
156
00:10:33,720 --> 00:10:38,280
ู„ูƒู„ ุงู„ู€ data setุŒ ูˆุฃุฐู‡ุจ ู„ุฃุฌู„ ุฃู† ุฃูˆุฌุฏุŒ ุฃูˆ ุฃุฌุฏ ู…ู†
157
00:10:38,280 --> 00:10:45,240
ุฎู„ุงู„ู‡ุง ุงู„ู€ root. ุงู„ุขู†ุŒ ูƒู„ ุงู„ู€ data set ุงู„ุชูŠ ุนู†ุฏูŠ ู‡ู†ุง
158
00:10:45,240 --> 00:10:52,780
must be categorical. ุงู„ุขู† ููŠ ุงู„ู€ C4.5ุŒ ูƒู„ ุงู„ู€
159
00:10:52,780 --> 00:10:56,500
attribute ู„ุงุฒู… ุชูƒูˆู† categorical. ุทูŠุจุŒ ุฃู†ุง ู…ุง ุนู†ุฏูŠุดุŒ ุฃู†ุง
160
00:10:56,500 --> 00:11:01,750
ุนู†ุฏูŠ continuous valueุŒ ุณุฃู‚ูˆู… ุจุนู…ู„ discretizationุŒ ูˆู„ูƒู„
161
00:11:01,750 --> 00:11:06,030
ุจุฅู…ูƒุงู†ูŠ ุฃู† ุฃูุนุทูŠู‡ุง labelุŒ ูˆุฃุนุชู…ุฏุŒ ุฃุดุชุบู„ ุนู„ู‰ ุงู„ู€ label ุงู„ู…ูˆุฌูˆุฏุฉ
162
00:11:06,030 --> 00:11:09,810
ุนู†ุฏูƒ ู‡ู†ุง. ูŠุนู†ูŠ ู„ู…ุง ุชุฃุชูŠ ู…ุซู„ุงู‹ ุงู„ู€ ageุŒ ู†ู‚ูˆู„ ูˆุงู„ู„ู‡
163
00:11:09,810 --> 00:11:19,390
youth ู…ู† 16 ู…ุซู„ุงู‹ ุฅู„ู‰ 22ุŒ youth. ุงู„ู€
164
00:11:19,390 --> 00:11:24,870
ageุŒ ุฃู‚ูˆู„ ู…ุซู„ุงู‹ ู…ู† 23 ุฅู„ู‰ 35ุŒ senior.
165
00:11:26,900 --> 00:11:30,060
ูˆุจุงู„ุชุงู„ูŠ ุฃู†ุง ุฃู‚ุฏุฑ ุฃุดุชุบู„... ุจู…ุง ุฃู†ู‘ ุงู„ู€ algorithm ูŠุฑูŠุฏ
166
00:11:30,060 --> 00:11:33,260
ู…ู†ูŠ discrete ุฃูˆ nominal dataุŒ ูุฃู‚ุฏุฑ ุฃุนู…ู„
167
00:11:33,260 --> 00:11:37,140
discretizationุŒ ุฃุนู…ู„ binningุŒ ูˆุจุนุฏ ุฐู„ูƒ ุฃุฐู‡ุจุŒ ุฃุถุน
168
00:11:37,140 --> 00:11:43,280
label ู„ูƒู„ bin ุฃูˆ ู„ูƒู„ interval ููŠ ุงู„ู€ continuous
169
00:11:43,280 --> 00:11:47,000
attribute ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ู…. ูˆุนู„ู‰ ุงู„ุฑุบู… ู…ู† ุฐู„ูƒุŒ ุณุฃุชูƒู„ู…
170
00:11:47,000 --> 00:11:51,160
ูƒุฐู„ูƒ ู‚ู„ูŠู„ู‹ุงุŒ ุจุดูƒู„ ุจุณูŠุทุŒ ููŠ ู…ูˆุถูˆุนุŒ ูุนู„ูŠู‹ุง ูƒูŠู ุฃู…ูƒู†ู†ูŠ
171
00:11:51,160 --> 00:11:54,620
ุฃูุญุต ู„ูˆ ูƒุงู† ุนู†ุฏูŠ continuous. ููŠ algorithms ู…ุฎุชู„ูุฉ.
172
00:11:54,620 --> 00:11:55,680
ุทูŠุจ.
173
00:11:58,590 --> 00:12:03,350
ู…ู…ุชุงุฒ. ู…ุนู†ุงุชู‡ ุฃู†ุง ูุนู„ูŠู‹ุง ุณุขุฎุฐ ุงู„ู€ data setุŒ ูˆุฃุจุฏุฃ ุฃู…ุฑุฑ
174
00:12:03,350 --> 00:12:08,530
ุนู„ู‰ ูƒู„ attribute ูˆุนู„ู‰ ูƒู„ ุงู„ู€ rowsุŒ ูˆุฃูุฌุฒู‘ุฆ ุงู„ุนู†ุงุตุฑ
175
00:12:0
223
00:15:57,930 --> 00:16:01,530
student ููƒุงู†ุช ู‡ูŠ ุงู„ู€ student ุงู„ู€ student ู‡ุฏููŠ yes ูˆ
224
00:16:01,530 --> 00:16:06,370
no ุจู†ุงุก ุนู„ูŠู‡ุŒ ุงู„ู€ data set ุจุชู†ุฌูŠ ุณูŠู…ู„ุฉ two data sets
225
00:16:06,370 --> 00:16:10,450
ูƒู…ุงู† ู…ุฑุฉ ูˆุงุญุฏุฉ ู…ุน ุงู„ู€ yes ูˆ ูˆุงุญุฏุฉ ู…ุน ุงู„ู€ noุŒ ูˆ ุจู†ู‚ู„
226
00:16:10,450 --> 00:16:14,510
ุทุจุนุง ุจู…ุง ุฃู†ู‡ ุฃู†ุง student yes ูˆ no ุจุฑูˆุญ ุจุฏูˆุฑ ุตุงุฑุช
227
00:16:14,510 --> 00:16:17,450
ู‡ุฏูˆู„ุฉ ุจูŠู†ุชู…ูˆุง ู„ุฏูˆู„ุฉ class ูˆ ู‡ุฏูˆู„ุฉ ูƒู„ partition ุจูŠู†ุชู…ูˆุง ู„ู€
228
00:16:17,450 --> 00:16:22,100
ุงู„ู€ class ู…ุนู†ุงุชู‡ ุฃู†ุง ูˆุฌูุชุŒ ุทูŠุจ ูุญุตุช ุงู„ู€ .. ุนููˆุงุŒ ูุญุตุช
229
00:16:22,100 --> 00:16:26,320
ุงู„ู€ age ูˆ ูุญุตุช ุงู„ู€ student ูˆ ูุญุตุช ุงู„ู€ income ูˆ ููŠ
230
00:16:26,320 --> 00:16:28,940
ุงู„ุขุฎุฑุŒ ู„ุงุฌูŠุช ุฃู† ุฃู†ุง ูุนู„ูŠุง ู…ุง ููŠุด ุนู†ุฏูŠ attributes
231
00:16:28,940 --> 00:16:32,200
ูุฎู„ุตู†ุงุŒ ูู‡ุฐู‡ ุงู„ู€ condition ุฃูˆ stopping conditions
232
00:16:32,200 --> 00:16:35,900
ุงู„ู„ูŠ ุฃู†ุง ู…ู…ูƒู† ุฃูˆู‚ู ุนู„ูŠู‡ุง ุทุงู„ู…ุง ุงู„ู€ data ุณุชุฉ ุจู‚ุช
233
00:16:35,900 --> 00:16:40,840
ูƒุจูŠุฑุฉ ูˆ ููŠู‡ุง ุดุบู„ุŒ ุงู„ู€ decision tree ุจูŠุงุฎุฏ ู…ู†ูŠ ูˆู‚ุช ุทูˆูŠู„ ููŠ
234
00:16:40,840 --> 00:16:44,820
ู…ูˆุถูˆุน ุงู„ู‚ุฑุงุฑ ุงู„ู„ูŠ ู…ูˆุฌูˆุฏ ู‡ู†ุงุŒ ูƒู…ุงู† ู…ุฑุฉ ุจู„ุฎุต ุจุดูƒู„
235
00:16:44,820 --> 00:16:50,430
ุณุฑูŠุนุŒ ู…ุชู‰ ุฃู†ุง ู…ู…ูƒู† ุฃูˆู‚ูุŒ ุฃุธู„ ุฃุจุญุซุŒ ู…ูŠู† ุงู„ู„ูŠ ุจูŠุญุฏุฏ ุงู„ู€
236
00:16:50,430 --> 00:16:53,630
depth ุชุจุนุช letter EุŒ ุงู„ู€ depth ุชุจุนุช letter E ุชุญุฏุฏ
237
00:16:53,630 --> 00:16:56,810
ุชุจุน ุงู„ู€ dimensionality ุชุจุน ุงู„ู€ data setุŒ ุนุฏุฏ ุงู„ู€
238
00:16:56,810 --> 00:17:01,550
attributes ูˆ ุนุฏุฏ ุงู„ู€ rowsุŒ ุงู„ุขู† ู…ุชู‰ ุจุฏูŠ ุฃูˆู‚ูุŸ ู„ู…ุง
239
00:17:01,550 --> 00:17:04,890
ุชูƒูˆู† ูƒู„ ุงู„ู€ sample ููŠ ุงู„ู€ given node ุชู†ุชู…ูŠ ู„ู†ูุณ ุงู„ู€
240
00:17:04,890 --> 00:17:07,950
classุŒ ูŠุนู†ูŠ ู„ู…ุง ุฃู†ุง ุงุฎุชุฑุช ุงู„ู€ attribute ูˆ ุฑูˆุญุช ุฃุนู…ู„
241
00:17:07,950 --> 00:17:11,970
splitุŒ ู„ุงุฌูŠุช ูˆุงุญุฏุฉ ู…ู† ุงู„ู€ partitions ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ
242
00:17:11,970 --> 00:17:13,970
ุจุชู†ุชู…ูŠ ู„ู†ูุณ ุงู„ู€ classุŒ ุฎู„ุงุตุŒ ู‡ุงุฏ ุงุจู†ูŠุŒ ู…ุง ููŠุด ููŠู‡ุง ุดุบู„
243
00:17:13,970 --> 00:17:16,870
ู„ุฅู† ู‡ุงุฏ ุงู„ู€ decision ุชุจุนุชู‡ุง ู…ุจุงุดุฑุฉุŒ ู„ุฅู†ู‡ุง ุจุชู†ุชู…ูŠ
244
00:17:16,870 --> 00:17:20,850
ู„ู†ูุณ ุงู„ู€ classุŒ ู…ุนู†ุงุชู‡ ุฎู„ุงุต ูˆุตู„ุช ู„ู€ leave nodeุŒ ุงู„ู€
245
00:17:20,850 --> 00:17:29,750
leave node ุชุจุนุชูŠุŒ ุงู„ู€ leave nodeุŒ ุงู„ู€ leaveุŒ ู„ู…ุง ุงู„ุฎูŠุงุฑ
246
00:17:29,750 --> 00:17:32,690
ุงู„ุชุงู†ูŠุŒ ุฃู†ู‡ ู„ู…ุง ุจูƒูˆู† ูุนู„ูŠุง ุฃู†ุง ุนู…ู„ุช splitting ู„ู„ู€
247
00:17:32,690 --> 00:17:35,250
data set ุนู„ู‰ ูƒู„ ุงู„ู€ attributes ูˆ ุฎู„ุตุช ุงู„ู€ attributes
248
00:17:35,250 --> 00:17:40,030
ุชุจุนุชูŠุŒ ุจุฑุถู‡ ู…ุง ุนู†ุฏูŠุด ุดุบู„ุŒ ูˆ there is no sample left
249
00:17:40,030 --> 00:17:43,610
ู…ุง ุจู‚ุงุด ุนู†ุฏูŠ ูˆู„ุง ุญุงุฌุฉ ููŠ ุงู„ู€ data set ุนุดุงู† ุฃุฌุณู…ู‡ุง
250
00:17:43,610 --> 00:17:47,510
ุนู„ู‰ ู…ุณุชูˆู‰ ุงู„ุฑุฃุณุŒ ุชุนุงู„ูˆุง ู†ุฑูˆุญ ู…ุน ุจุนุถ ู…ู† ุฎู„ุงู„ ู†ุดูˆู ุงู„ู€
251
00:17:47,510 --> 00:17:52,910
information gain ูˆ ู‡ูŠ ุงู„ุฃูƒุซุฑ ูˆ ุงู„ุฃุดู‡ุฑ ุงุณุชุฎุฏุงู…ุงุŒ ูˆ
252
00:17:52,910 --> 00:17:55,830
ุงู„ู€ Gain IndexุŒ ู‡ู†ุชูƒู„ู… ุนู„ู‰ ุงู„ู€ information gain ุจูƒู„
253
00:17:55,830 --> 00:17:59,970
ุจุณุงุทุฉุŒ ุงู„ู€ information gain ุจุชุนุชู…ุฏ ุนู„ู‰ ุงู„ู€
254
00:17:59,970 --> 00:18:02,830
probabilityุŒ ู…ุด ุฅุญู†ุง ู‚ู„ู†ุง ู‚ุจู„ ุดูˆูŠุฉ ู…ูˆุถูˆุน ุงู„ู€
255
00:18:02,830 --> 00:18:06,970
splitting ุฃูˆ ุงู„ูุตู„ ููŠ ุงู„ู€ attributes ุจูŠุนุชู…ุฏ ุงุนุชู…ุงุฏุง
256
00:18:06,970 --> 00:18:12,250
ูƒู„ู‡ ุนู„ู‰ ูุนู„ูŠุง ุงู„ุนู†ุงุตุฑ ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉุŒ ุนู„ู‰ ุงุญุชู…ุงู„
257
00:18:12,250 --> 00:18:15,930
ูŠู‚ูˆู„ ู„ู†ุง ุฅู…ุง heuristic rules ุฃูˆ statistical
258
00:18:15,930 --> 00:18:19,650
measurementุŒ ู„ู…ุง ุจุชูƒู„ู… ุนู„ู‰ probability ู…ุนู†ุงุชู‡ ุฃู†ุง
259
00:18:19,650 --> 00:18:22,950
ุฌุงูŠ ุจุชูƒู„ู… ุนู„ู‰ ุงุญุชู…ุงู„ุงุชุŒ ุงู„ู€ statistics ุฅู„ู‰ ุขุฎุฑู‡
260
00:18:23,780 --> 00:18:27,160
ุจู‚ูˆู„ุŒ ุงูุชุฑุถ ุฃู† ุงู„ู€ P I ู‡ูŠ ุนุจุงุฑุฉ ุนู† ุงู„ู€ probability
261
00:18:27,160 --> 00:18:34,780
of an arbitrary tuple ููŠ ุงู„ู€ data ุงู„ู€ 6 ุชุจุนุชูŠ
262
00:18:34,780 --> 00:18:36,740
ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ
263
00:18:36,740 --> 00:18:36,820
ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ
264
00:18:36,820 --> 00:18:37,520
ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ
265
00:18:37,520 --> 00:18:40,280
ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ
266
00:18:40,280 --> 00:18:49,560
ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุนุชูŠ ุชุจุน
267
00:18:51,740 --> 00:18:55,960
ุงู„ู€ Probability ู„ู„ู€ CDุŒ ุงู„ู€ CI ุนู„ู‰ ุงู„ู€ DุŒ ุนู„ู‰ ูƒู„
268
00:18:55,960 --> 00:18:59,300
Probability ุชุจุน ุงู„ู€ data set ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉุŒ ุจูƒู„ ุจุณุงุทุฉ
269
00:18:59,300 --> 00:19:04,400
ุฃู†ุง ููŠ ุนู†ุฏูŠ ุชู„ุช ุนู…ู„ูŠุงุช ุญุณุงุจูŠุฉ ู‡ุนู…ู„ู‡ุง ุนุดุงู† ุขุฎุฏ ุงู„ู€
270
00:19:04,400 --> 00:19:07,560
decision ูˆ ุฃุญุฏุฏ ู…ู† ุงู„ู€ test node ุงู„ู„ูŠ ู…ูˆุฌูˆุฏ ุนู†ุฏูŠ
271
00:19:07,560 --> 00:19:11,540
ุงู„ุฃูˆู„ู‰ุŒ ู‡ุณู…ูŠู‡ุง ุงู„ู€ expected information ุฃูˆ ุงู„ู€
272
00:19:11,540 --> 00:19:19,260
entropyุŒ ูˆู‡ูŠ ู„ูƒู„ ุงู„ู€ data set ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ู‡ู†ุง
273
00:19:19,260 --> 00:19:27,600
ุดูˆ ูŠุนู†ูŠุŸ ุงู„ุขู† ู…ุทู„ูˆุจ ู…ู†ูŠ ุฃู† ุฃุญุณุจ ุงู„ู€ information ุฃูˆ
274
00:19:27,600 --> 00:19:31,140
ุงู„ู€ entropy ู„ู„ู€ classes ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ููŠ ุงู„ู€ data set
275
00:19:31,140 --> 00:19:34,700
ู„ูƒู„ ุงู„ู€ data setุŒ ูˆู‡ู†ุง ุจู†ุฌูˆ ุณูŠู†ุŒ ูˆูƒุฃู†ูŠ ุจุฏู‡ ูŠู‚ูˆู„ ู„ูŠ
276
00:19:34,700 --> 00:19:40,940
ุงุญุณุจ ุงุญุชู…ุงู„ูŠุฉุŒ ุฃูˆ ุงุญุณุจ ุงู„ู€ probability ู„ูƒู„ class ููŠ
277
00:19:40,940 --> 00:19:43,880
ุงู„ู€ data setุŒ ุนุฏุฏ ู…ุฑุงุช ุธู‡ูˆุฑ ุงู„ู€ class ููŠ ุงู„ู€ data set
278
00:19:43,880 --> 00:19:46,840
ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ู‡ู†ุงุŒ ูˆุจุนุฏ ู‡ูŠูƒ ุจุฑูˆุญ ุฃุทุจู‚ ุนู„ูŠู‡ู…
279
00:19:46,840 --> 00:19:51,840
ุงู„ุนู…ู„ูŠุฉุŒ ูŠุนู†ูŠ ุฃู†ุง ู„ูˆ ูƒู†ุช ุนู„ู‰ ุณุจูŠู„ ุงู„ู…ุซุงู„ุŒ ุงู„ู€ data 6
280
00:19:51,840 --> 00:19:57,580
ุชุจุนุชูŠ ููŠู‡ุง ุนุดุฑ elementsุŒ ุฃุฑุจุนุฉ ู…ู†ู‡ู… yes ูˆุณุชุฉ ู…ู†ู‡ู… no
281
00:19:57,580 --> 00:20:01,040
ุฃู†ุง
282
00:20:01,040 --> 00:20:04,100
ููŠ ุนูŠู†ูŠ ุจุชูƒู„ู… ุนู„ู‰ binary classification yes or no
283
00:20:04,100 --> 00:20:10,200
ุฃุฑุจุนุฉ yes ูˆุณุชุฉ noุŒ ุงู„ู€ information gained ู„ู„ู€ data 6
284
00:20:10,200 --> 00:20:16,720
ุชุจุนุชูŠ ูƒู„ู‡ุง ู‡ูŠ ุนุจุงุฑุฉ ุนู† ู…ุฌู…ูˆุน ู†ุงู‚ุต
285
00:20:16,720 --> 00:20:20,600
ู…ุถุฑูˆุจุฉ ููŠู‡ ุงู„ู€ probability ุชุจุนุช ุงู„ู€ data ุงู„ุฃูˆู„ู‰
286
00:20:22,150 --> 00:20:28,310
ุงุญุชู…ุงู„ ุงู„ู€ class ุงู„ุฃูˆู„ 4 ุนู„ู‰ 10 ู…ุถุฑูˆุจุฉ ููŠ log ุงู„ู€ 4
287
00:20:28,310 --> 00:20:33,210
ุนู„ู‰ 10 ู„ู„ุฃุณุงุณ 2ุŒ ุฒุงุฆุฏ
288
00:20:34,590 --> 00:20:40,310
6 ุนู„ู‰ 10 ู…ุถุฑูˆุจุฉ ููŠ ุงู„ู€ logarithm 6 ุนู„ู‰ 10ุŒ ุงู„ู€
289
00:20:40,310 --> 00:20:44,350
logarithmุŒ ุงู„ู€ binary logarithmุŒ ูˆู‡ูƒุฐุงุŒ ู„ูŠุด ุงู„ู‚ูŠู…ุฉ
290
00:20:44,350 --> 00:20:48,030
ู‡ุชุฏุงู†ูŠ ุฅุดุงุฑุฉ ุณุงู„ุจุฉุŸ ู„ุฃู† ุงู„ู€ logarithm ุชุจุนุช ุงู„ู€ binary
291
00:20:48,030 --> 00:20:52,690
ุจุชุงุนุฉ ุงู„ูƒุณู„ ู‡ุชุทู„ุน ุนู†ุฏูŠ ุณุงู„ุจุŒ ูˆุฅุฐุง ุจุชุฐูƒุฑูˆุงุŒ ููŠ ุนู†ุฏูŠ
292
00:20:52,690 --> 00:21:01,740
log ุงู„ู€ X ุนู„ู‰ ุงู„ู€ Y ุชุณุงูˆูŠ log X ู†ุงู‚ุต log YุŒ ูˆุจู…ุง ุฃู† ุงู„ู€
293
00:21:01,740 --> 00:21:05,200
Y ุนู†ุฏูŠ ุฃูƒุจุฑ ู…ู† ุงู„ู€ X ูุณุชูƒูˆู† ุงู„ู‚ูŠู…ุฉ ุงู„ู„ูŠ ุนู†ุฏูŠ ุณุงู„ุจุฉ
294
00:21:05,200 --> 00:21:07,880
ุนุดุงู† ุฃู†ุง ุฃุฎู„ุต ู…ู†ู‡ุงุŒ ุฃุฎู„ุต ู…ู†ู‡ุงุŒ ููƒุงู†ุช ุงู„ู‚ูŠู…ุฉ ุงู„ู„ูŠ
295
00:21:07,880 --> 00:21:12,360
ุนู†ุฏูŠ ู‡ู†ุง ู‡ุชุทู„ุน ู‚ูŠู… ู…ูˆุฌุจุฉุŒ ูˆุงุถุญ ุงู„ุฃู…ูˆุฑ ุฅู† ุดุงุก ุงู„ู„ู‡
296
00:21:12,360 --> 00:21:18,090
ูŠุง ุฌู…ุงุนุฉ ุงู„ุฎูŠุฑุŒ ุงู„ุขู† ูŠุนู†ูŠ ุฃูˆู„ ุดุบู„ุฉ ูุนู„ูŠุง ุฃู†ุง ู‡ุณูˆูŠู‡ุง
297
00:21:18,090 --> 00:21:21,350
ู‡ุฑูˆุญ ุฃุญุณุจ ุงู„ู€ probability ู„ูƒู„ class ุฃูˆ ุจูŠู† ุฌุซูŠู†
298
00:21:21,350 --> 00:21:27,670
ู‡ุญุณุจ ุงู„ู€ entropyุŒ ู‡ูŠ ุนุจุงุฑุฉ ุนู† ุญุงุตู„ .. ุนุจุงุฑุฉ ุนู† ู…ุฌู…ูˆุน
299
00:21:27,670 --> 00:21:33,310
ุญุงุตู„ ุถุฑุจ ุงู„ู€ probability ู„ูƒู„ class ููŠ ุงู„ู€ logarithm
300
00:21:33,310 --> 00:21:38,830
ุฃูˆ ุงู„ู€ binary log ู„ู€ ุงู„ู€ probability ู„ู€ ุงู„ู€ class ุงู„ู„ูŠ
301
00:21:38,830 --> 00:21:41,930
ู…ูˆุฌูˆุฏ ุนู†ุฏูŠุŒ ูˆ ุฒูŠ ู…ุง ุจู‚ูˆู„ู‡ ุจุงู„ู€ .. ุจุงู„ู€ .. ุจุงู„ู€ ..
302
00:21:42,360 --> 00:21:45,340
ุจุงู„ู…ุซุงู„ ูŠุชุถุญุŒ ูƒู…ุงู† ุดูˆูŠุฉ ู‡ู†ุชู‚ู„ ู„ู„ู…ุซุงู„
303
00:21:45,340 --> 00:21:50,740
ุจุงู„ุชูุตูŠู„ ุฅู† ุดุงุก ุงู„ู„ู‡ ุชุนุงู„ู‰ุŒ ุงู„ุฎุทูˆุฉ ุงู„ู„ูŠ ุจุนุฏ ู‡ูŠูƒ ุจุฏูŠ
304
00:21:50,740 --> 00:21:58,260
ุฃุฑูˆุญ ู„ูƒู„ attribute AุŒ ุฃุญุงูˆู„ ูุนู„ูŠุง ุฃุฌุณู… ุงู„ู€ data set
305
00:21:58,260 --> 00:22:02,040
ู„ู…ุฌู…ูˆุนุฉ ู…ู† ุงู„ู€ partitionsุŒ ุฌุฏุงุด ุนุฏุฏ ุงู„ู€ V ู‡ุฐู‡ุŒ ุฃูˆ ุฌุฏุงุด
306
00:22:02,040 --> 00:22:07,420
ุนุฏุฏ ุงู„ู€ partitionsุŒ ุจุนุฏุฏ ุงู„ู€ distinct values ุงู„ู„ูŠ
307
00:22:07,420 --> 00:22:12,630
ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ููŠ ุงู„ู€ attributeุŒ ูŠุนู†ูŠ ุงู„ู€ age ูƒุงู† ููŠ
308
00:22:12,630 --> 00:22:19,910
ุนู†ุฏูŠ ุชู„ุงุชุฉุŒ three valuesุŒ ูŠุซ ูˆ middle age ูˆ
309
00:22:19,910 --> 00:22:23,570
seniorุŒ ููุนู„ูŠุง ุงู„ู€ attribute ุงู„ู€ age ู‡ูŠ
310
00:22:26,960 --> 00:22:30,380
ุงู„ู€ Attribute ุงู„ู€ AgeุŒ ูŠุซ ูˆ Middle Age ูˆ senior
311
00:22:30,380 --> 00:22:33,860
ู‡ูŠ ุงู„ู€ three distinct valuesุŒ ููุนู„ูŠุง ุจู†ุงุก ุนู„ู‰ ุงู„ู€
312
00:22:33,860 --> 00:22:38,600
attribute ู‡ุฐุงุŒ ู‡ุฌุณู… ุงู„ู€ data set ุชุจุนุชูŠ ูƒู„ู‡ุง ู„ู€ three
313
00:22:38,600 --> 00:22:43,860
.. ู„ู€ three subsetsุŒ ู„ู€ three partitionsุŒ ู…ุน ูƒู„ ูˆุงุญุฏุฉ
314
00:22:43,860 --> 00:22:46,600
ู…ู† ุงู„ู€ values ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ู‡ู†ุงุŒ ุนุดุงู† ุฃู†ุง ูุนู„ูŠุง
315
00:22:46,600 --> 00:22:52,190
ุฃุฑูˆุญ ุฃุญุณุจ ุงู„ู€ information ู„ู„ู€ attribute ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ
316
00:22:52,190 --> 00:22:56,050
ุนู†ุฏู‡ุงุŒ ุฅูŠุด ุงู„ู€ information ู„ู„ู€ attributeุŸ ุงู„ู€
317
00:22:56,050 --> 00:23:01,870
information ู„ู„ู€ attribute ูŠุณุงูˆูŠ ุงู„ู€ summationุŒ ุงุญุชู…ุงู„
318
00:23:01,870 --> 00:23:04,870
ุงู„ู€ element ุงู„ู„ูŠ ุนู†ุฏูŠุŒ ุฃูˆ ุงู„ู€ class ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ
319
00:23:04,870 --> 00:23:08,310
ู‡ู†ุง ู„ู„ู€ attributeุŒ ููŠู‡ ุงู„ู€ information ุชุจุนุช ุงู„ู€ subset
320
00:23:08,310 --> 00:23:12,390
ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉุŒ ูŠุนู†ูŠ ุจูŠู† ุฌุณูŠู… ุงู„ู€ subset ุงู„ุฌุฏูŠุฏุฉุŒ ุฃู†ุง
321
00:23:12,390 --> 00:23:16,770
ู‡ุดุชุบู„ ุนู„ูŠู‡ุงุŒ ูˆ ุฃุญุณุจู‡ุงุŒ ุงู„ู€ intro b ุงู„ุณุงุจู‚ุฉุŒ ุจุนุฏ ู…ุง
322
00:23:16,770 --> 00:23:23,570
ุฃุญุณุจ ุงู„ู€ information ู„ู„ู€ attributeุŒ ุงู„ู€ gainุŒ ุงู„ุงู†ุญูŠุงุฒ
323
00:23:25,400 --> 00:23:28,740
ู„ู€ ุงู„ู€ element ุงู„ู„ูŠ ู…ูˆุฌูˆุฏ ุนู†ุฏูŠ ู‡ู†ุงุŒ ุฃูˆ ุงู„ุชุญุตูŠู„ ุงู„ู„ูŠ
324
00:23:28,740 --> 00:23:33,400
ู…ู…ูƒู† ู†ุณู…ูŠู‡ ุงู„ุชุญุตูŠู„ ุงู„ู€ aุŒ ู‡ูŠ ุนุจุงุฑุฉ ุนู† ุงู„ู€
325
00:23:33,400 --> 00:23:38,760
information ู„ู„ู€ D ุงู„ุฃูˆู„ู‰ุŒ ู†ุงู‚ุต ุงู„ู€ information ุฃูˆ ุงู„ู€
326
00:23:38,760 --> 00:23:41,640
entropy ุชุจุนุช ุงู„ู€ attribute ุงู„ู„ูŠ ุนู†ุฏูŠุŒ ุงู„ู€ entropy
327
00:23:41,640 --> 00:23:47,180
ุชุจุนุช ุงู„ู€ attribute ุงู„ู„ูŠ ุนู†ุฏูŠุŒ ูŠุนู†ูŠ ุฃู†ุง ูุนู„ูŠุง ู‡ุงุฎุฏ
328
00:23:47,180 --> 00:23:53,540
ู‡ุงูŠ ูˆ ู‡ุงูŠู‡ุทุฑุญู‡ู… ู…ู† ุจุนุถุŒ ุจุณ ู…ุง ุชู†ุณูˆุด ุฃู† ูุนู„ูŠุง ู‡ุฐู‡
329
00:23:53,540 --> 00:23:59,580
ู‡ูŠ ู†ูุณู‡ุง ุงู„ู„ูŠ ููˆู‚ุŒ ุจุณ ุนู„ู‰ different subset ุฃูˆ ุนู„ู‰
330
00:23:59,580 --> 00:24:03,680
different data set ุฃูˆ ุจูŠู† ุฌุณูŠู†ุŒ ุนู„ู‰ subset set ุชุจุนุง
331
00:24:03,680 --> 00:24:07,320
ู„ู„ู€ values ุงู„ู„ูŠ ุฌุณู…ุชู‡ุงุŒ ู„ู„ู€ partition ุงู„ู„ูŠ ุฌุณู…ุชู‡ุง ู„ูŠู‡ุง
332
00:24:07,320 --> 00:24:11,480
ุงู„ู€ attribute ุงู„ู„ูŠ ุนู†ุฏู†ุงุŒ ุชุนุงู„ูˆุง ู†ุฑูˆุญ ู‡ู†ุง ูˆ ู†ุดูˆู
333
00:24:11,480 --> 00:24:17,140
ุงู„ู…ุซุงู„ ุงู„ู„ูŠ ู…ูˆุฌูˆุฏ ุนู†ุฏู†ุงุŒ ู‚ู„ู†ุง ููŠ ุงู„ุฎุทูˆุฉ ุฑู‚ู… ูˆุงุญุฏุŒ ููŠ
334
00:24:17,140 --> 00:24:21,600
ุงู„ุฎุทูˆุฉ ุฑู‚ู… ูˆุงุญุฏุŒ ู‚ู„ู†ุง ุฃู†ุง ุงุชูู‚ู†ุง ุฃู† ุฃู†ุง ูุนู„ูŠุง ู‡ุฑูˆุญ
335
00:24:21,600 --> 00:24:24,800
ุฃุญุณุจ ุงู„ู€ information gain ุฃูˆ ุงู„ู€ entropy ู„ูƒู„ ุงู„ู€ data
336
00:24:24,800 --> 00:24:29,260
setุŒ ู…ู…ุชุงุฒุŒ ุนุดุงู† ุฃุญุณุจ ุงู„ู€ entropy ู„ูƒู„ ุงู„ู€ data set
337
00:24:29,260 --> 00:24:33,540
ุจู†ุฐูƒุฑุŒ ู‚ู„ู†ุง ู‡ูˆ ุนุจุงุฑุฉ ุนู† ุงู„ู€ summationุŒ ุณุงู„ุจ ูˆุงุญุฏ ููŠ
338
00:24:33,540 --> 00:24:40,160
ุงู„ู€ summation ููŠ probability ู„ู„ู€ I ููŠ log ู„ู€ ุงู„ู€ P I
339
00:24:40,160 --> 00:24:43,960
ุงู„ู€ probability ู„ู„ู€ IุŒ ูˆุฅูŠุด ู‚ู„ู†ุงุŸ ู‡ุงูŠ ุงู„ู€ classุŒ ูˆุงู„ู€ I
340
00:24:43,960 --> 00:24:46,920
ุจูŠุชุณุงูˆู‰ ู…ู† ูˆุงุญุฏ ู„ุนุฏุฏ ุงู„ู€ classes ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ
341
00:24:46,920 --> 00:24:51,360
ุญุณุจ ุงู„ู€ data set ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ุงุŒ ุญุณุจ ุงู„ู€ data set
342
00:24:51,360 --> 00:24:53,900
ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ุงุŒ ุฃู†ุง ููŠ ุนู†ุฏูŠ two different
343
00:24:53,900 --> 00:24:58,680
classesุŒ only two different classes onlyุŒ ุงู„ู„ูŠ ู‡ู…
344
00:24:58,680 --> 00:25:05,940
yes ูˆ noุŒ ุญุฌู… ุงู„ู€ data set ูƒู„ู‡ ุฌุฏุงุด ูŠุง ุฌู…ุงุนุฉ ุงู„ุฎูŠุฑุŸ 14
345
00:25:05,940 --> 00:25:11,280
ุนุฏุฏ ุงู„ู€ yes 9ุŒ ู…ุนู†ุงุชู‡ ุงู„ู€ probability ุชุจุนุชู‡ุง 9 ุนู„ู‰ 14
346
00:25:13,570 --> 00:25:18,130
ุงู„ู€ Probability ู„ู„ู€ Yes 9 ุนู„ู‰ 14ุŒ ุทุจ ุนุฏุฏ ุงู„ู€ No
347
00:25:18,130 --> 00:25:24,810
ุจุฌูŠุชู‡ุง 5ุŒ 5 ุนู„ู‰ 14ุŒ ู‡ูŠ ุงู„ุงุญุชู…ุงู„ูŠุฉ ุชุจุนุชู‡ุง 5 ุนู„ู‰ 14
348
00:25:24,810 --> 00:25:31,190
ูˆุจุงู„ู‡ูŠูƒ ุฃู†ุง ุญุตู„ุช ุนู„ู‰ ุฃูˆู„ ุฎุทูˆุฉุŒ ุงู„ู„ูŠ ู‡ูŠ ุนุฑูุช ุงู„ู€
349
00:25:31,190 --> 00:25:34,790
probability ุชุจุนุช ู„ู€ classes ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠุŒ ุงู„ู„ูŠ
350
00:25:34,790 --> 00:25:37,570
ู‡ูŠ ุงู„ู€ probability ู„ู„ู€ yes ูˆ ุงู„ู€ probability ู„ู„ู€ no
351
00:25:37,570 --> 00:25:40,030
ุงู„ุฎุทูˆุฉ ุงู„ุชุงู„ูŠุฉ ุงู„ู„ูŠ ุงู„ู…ูุฑูˆุถ ุฃู† ุฃู†ุง ุฃุดุชุบู„ ุนู„ูŠู‡ุง
352
00:25:40,030 --> 00:25:43,410
ุงู„ุขู†ุŒ ุฃู† ุฃู†ุง ุจุฏูŠ ุฃุฑูˆุญ ุฃุญุณุจ ุงู„ู€ information ุฃูˆ ุงู„ู€
353
00:25:43,410 --> 00:25:48,210
entropy ุชุจุนุชูŠุŒ ุจุงู„ู…ุนุงุฏู„ุฉ ุงู„ุชุงู„ูŠุฉุŒ ุงู„ู€ information
354
00:25:48,210 --> 00:25:57,610
ู‡ู†ุฑู…ุฒู„ู‡ุง ู„ู„ู€ I ู„ู„ู€ data set ุชุจุนุชูŠุŒ ุชุณุงูˆูŠ ุชุณุงูˆูŠ
355
00:25:59,800 --> 00:26:04,940
IุŒ ุชุณุนุฉ ูˆูƒู…ุง ุฎู…ุณุฉุŒ ุชุณุนุฉ ูˆุฎู…ุณุฉุŒ ู‡ุฏูˆู„ ู‡ู… ุงู„ุฃุฑุจุนุฉ ุนุดุฑ
356
00:26:04,940 --> 00:26:07,660
ุชุจุนูˆุชูŠ ูŠุง ุฌู…ุงุนุฉ ุงู„ุฎูŠุฑุŒ ุงู„ุขู†ุŒ ู‡ูŠ ุงู„ู‚ุงู†ูˆู† ุชุจุน ุงู„ู€
357
00:26:07,660 --> 00:26:12,580
informationุŒ ู‡ูŠู‡ ุจูŠู„ุฒู…ู†ูŠ ููŠู‡ุง ุฃู† ุฃุนุฑู ุงู„ู€
358
00:26:12,580 --> 00:26:16,020
probability ู„ู„ู€ yes ูˆ ุงู„ู€ probability ู„ู„ู€ noุŒ ูˆู‡ุฐุง
359
00:26:16,020 --> 00:26:24,260
ุงู„ูƒู„ุงู… ูŠุณุงูˆูŠุŒ ู…ุง ู‚ุต ู…ุถุฑูˆุจุฉ ููŠู‡ ูƒุงู… classุŸ and two
360
00:26:24,260 --> 00:26:27,790
classุŒ ู‡ูŠู‡ู… ู‡ุฏูˆู„ุŒ ู„ูˆ ูƒุงู†ูˆุง ุชู„ุงุชุฉ ู‡ูŠูƒูˆู†ูˆุง ุชู„ุงุชุฉุŒ ู„ูˆ
361
00:26:27,790 --> 00:26:32,130
ุฃุฑุจุนุฉ ู‡ูŠูƒูˆู†ูˆุง ุฃุฑุจุนุฉุŒ ุฅู„ู‰ ุขุฎุฑู‡ุŒ ู…ุน ูƒู„ ูˆุงุญุฏุฉ ููŠู‡ู…ุŒ ุงู„ุขู†
362
00:26:32,130 --> 00:26:36,270
ุชุณุนุฉ ุนู„ู‰ ุฃุฑุจุนุฉ ุนุดุฑุŒ ู‡ูŠ ุชุจุนุช ุงู„ู€ class ุงู„ุฃูˆู„ ุงู„ู„ูŠ
363
00:26:36,270 --> 00:26:42,030
ุจูŠู† ุฌุซูŠู†ุŒ ุฅุญู†ุง ู‚ู„ู†ุง ุงู„ู€ yes ู…ุถุฑูˆุจุฉ ููŠ ุงู„ู€ binary
364
00:26:42,030 --> 00:26:48,970
logarithm ู„ู„ุชุณุนุฉ ุนู„ู‰ ุฃุฑุจุนุฉ ุนุดุฑุŒ ู…ุฌู…ูˆุนุฉ ู„ู‡ู… ู…ุฌู…ูˆุนุฉ
365
00:26:48,970 --> 00:26:54,010
ุงู„ุฎู…ุณุฉ ุชุจุนุชู‡ุงุŒ ุงู„ุฎู…ุณุฉ ุชุจุนุช ุงู„ู€ noุŒ ุฎู…ุณุฉ ุนู„ู‰ ุฃุฑุจุนุฉ ุนุดุฑ
366
00:26:54,010 --> 00:27:01,670
ู…ุถุฑูˆุจุฉ ููŠ ุงู„ู€ logarithmุŒ ุงู„ู€ binary logarithmุŒ ุงู„ุฎู…ุณุฉ
367
00:27:01,670 --> 00:27:07,950
ุนู„ู‰ ุฃุฑุจุนุฉ ุนุดุฑุŒ ู‡ุฐุง ุงู„ู€ gain ุชุจุนุช ูƒู„ ุงู„ู€ data set
368
00:27:07,950 --> 00:27:11,230
ุชูˆุฒูŠุนุช ุงู„ู€ data set ุนู†ุฏูŠ ุนู„ู‰ two classesุŒ ุชุฐูƒุฑ ูƒู…ุงู†
369
00:27:11,230 --> 00:27:17,750
ู…ุฑุฉ ูˆุงุญุฏุฉุŒ ุงุชู†ูŠู†ุŒ ุชู„ุงุชุฉุŒ ุฃุฑุจุนุฉุŒ ุฎู…ุณุฉ noุŒ ุฎู…ุณุฉ ู…ู† ุฃุฑุจุนุฉ
370
00:27:17,750 --> 00:27:21,290
ุนุดุฑุŒ ู…ุนู†ุงุชู‡ ุนู†ุฏูŠ ุชุณุนุฉ yesุŒ ูˆุงู„ุขู† ู‚ูˆู„ู†ุงุŒ ู‡ูŠ ู‚ุงู†ูˆู†ู‡ู…
371
00:27:21,290 --> 00:27:25,670
ู‚ุงู†ูˆู† ุงู„ู€ information ุฃูˆ ุงู„ู€ gainุŒ ุนููˆุงุŒ ุงู„ู€ entropy ู„ู„ู€
372
00:27:25,670 --> 00:27:30,730
data set ู‡ูŠ ุนุจุงุฑุฉ ุนู† ุณุงู„ุจ ูˆุงุญุฏ ู…ุถุฑูˆุจุฉ ููŠ ู…ุฌู…ูˆุน ู„ู€
373
00:27:30,730 --> 00:27:35,410
probability ู„ูƒู„ class ู…ุถุฑูˆุจุฉ ููŠ ุงู„ู€ log ู„ู€ log ู„ู€
374
00:27:35,410 --> 00:27:38,650
probability ู„ูƒู„ classุŒ ูุฃู†ุง ู‡ู†ุง ุญุณุจุช ุงู„ู…ุนุงุฏู„ุฉ ุงู„ู„ูŠ
375
00:27:38,650 --> 00:27:40,670
ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠุŒ ู‡ุชุธู‡ุฑ ู…ู† ุฎู„ุงู„
376
00:27:43,770 --> 00:27:51,650
ุณู„ุงูŠุฏ ู‡ู†ุงุŒ ูˆู‡ุฐู‡
377
00:27:51,650 --> 00:27:59,370
ู‚ูŠู…ุชู‡ุง 0.94 ุฃูˆ ุฃุฑุจุนุฉ ูˆุชุณุนูŠู† ู…ู† ู…ูŠุฉุŒ ู‡ุฐู‡
378
00:27:59,370 --> 00:28:03,130
ุซุงุจุชุฉ ู‡ุชูƒูˆู† ู„ูƒู„ ุงู„ู€ data setุŒ ู„ูƒู„ training set ุงู„ู„ูŠ
379
00:28:03,130 --> 00:28:06,830
ุฃู†ุง ุจุจู†ูŠ ุนู„ูŠู‡ุง ุงู„ู€ modelุŒ ู…ู…ุชุงุฒุŒ ุฃู†ุง ู…ุดูŠุช ุฃูˆู„ ุฎุทูˆุฉ ููŠ
380
00:28:06,830 --> 00:28:12,030
ุงู„ุญู„ุŒ ุงู„ุฎุทูˆุฉ ุงู„ุชุงู„ูŠุฉ ุฃู† ุฃู†ุง ูุนู„ูŠุง ุจุฏูŠ ุฃุฑูˆุญุŒ ุฃูŠ ุงุชู†ูŠู†
381
00:28:16,020 --> 00:28:22,480
ุฃุจุฏุฃ ุฃุญุณุจ ุงู„ู€ information ู„ูƒู„ attribute ู…ูˆุฌูˆุฏ ููŠ ุงู„ู€
382
00:28:22,480 --> 00:28:28,200
data setุŒ ู„ูƒู„ attributeุŸ ุตุญูŠุญุŒ ูู‡ุฃุฎุฏ ุฃุญุณุจ ุงู„ู€ intro
383
00:28:28,200 --> 00:28:33,140
ุจุงู„ุฅู† ุฃูˆ ุงู„ู€ information gain ู„ู„ู€ ageุŒ ูˆุฃุญุณุจ ุงู„ู€
384
00:28:33,140 --> 00:28:39,490
information Gain ู„ู„ู€ incomeุŒ ู„ู„ู€ studentุŒ ู„ู„ู€ credit
385
00:28:39,490 --> 00:28:45,670
ratingุŒ ูˆู‡ูƒุฐุงุŒ ุฎู„ูŠู†ูŠ ุฃู†ุง ุฃุจุฏุฃ ู…ุนุงูƒู… ูˆุฃุฐูƒุฑูƒู… ุฃู†
386
00:28:45,670 --> 00:28:48,670
ุฃู†ุง ูุนู„ูŠุง ุจุญุณุจ ุงู„ู€ information gain ู„ู„ู€ attribute
387
00:28:52,790 --> 00:28:56,270
ุชุจุน ุนู„ู‰ ุงู„ู€ data set ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ุงุŒ ู„ุฃู† ุงู„ู€
388
00:28:56,270 --> 00:28:59,750
summation ุนู„ู‰ ุนุฏุฏ ุงู„ู€ partitionsุŒ ุงู„ู€ summationุŒ ุงู„ู€ V
389
00:28:59,750 --> 00:29:02,830
ุฒูŠ ู…ุง ู‚ู„ู†ุง ู‚ุจู„ ุดูˆูŠุฉุŒ ู‡ูŠู‡ุงุŒ ุงู„ู„ูŠ ู‡ูŠ ุนุฏุฏ ุงู„ู€ partitions
390
00:29:02,830 --> 00:29:08,650
ุงู„ู„ูŠ ุนู†ุฏู‡ุงุŒ number of partitionsุŒ ุญุฌู…
391
00:29:08,650 --> 00:29:12,850
ุงู„ู€ partition ู„ู„ู€ data setุŒ ุนุฏุฏ ุนู†ุงุตุฑ ุงู„ู€ partition
392
00:29:12,850 --> 00:29:16,290
ุนู„ู‰ ุนุฏุฏ ุนู†ุงุตุฑ ุงู„ู€ data setุŒ ratioุŒ ุงุญุชู…ุงู„ ูˆู„ุง ู„ุฃุŸ
393
0
445
00:34:25,000 --> 00:34:30,520
ู„ู‡ุงูŠ ูˆู‡ุฏ ุงู„ู„ูŠ ุงู„ู…ูุฑูˆุถ ุชุณุงูˆูŠ ู†ุงู‚ุต ู…ุถุฑูˆุจุฉ ููŠู‡ ุซู„ุงุซุฉ
446
00:34:30,520 --> 00:34:37,040
ุนู„ู‰ ุฎู…ุณุฉ ููŠ ุงู„ู€ log ุฃูŠูˆุฉ ุฌุฏุงุด ููŠ ุงู„ู€ binary log ุตุญูŠุญ
447
00:34:37,040 --> 00:34:43,800
ุซู„ุงุซุฉ ุนู„ู‰ ุฎู…ุณุฉ ุฒุงุฆุฏ ุงุซู†ูŠู† ุนู„ู‰ ุฎู…ุณุฉ ููŠ ุงู„ู€ log ุงุซู†ูŠู† ุนู„ู‰
448
00:34:43,800 --> 00:34:49,600
ุฎู…ุณุฉ ู‡ุฐู‡ ุงู„ู€ intro ุจุงู„ุฃูˆู„ู‰ ู‡ุฐู‡ ุฌู…ุนุฉ ุงู„ุฎูŠุฑ ุตูุฑ ู„ูŠุด
449
00:34:49,600 --> 00:34:59,080
ู†ุงู‚ุต ุฃุฑุจุนุฉ ุนู„ู‰ ุฎู…ุณุฉ ุฃุฑุจุนุฉ ุนู„ู‰ ุฎู…ุณุฉ ุฃุฑุจุนุฉ ุนู„ู‰ ุฎู…ุณุฉ
450
00:34:59,080 --> 00:35:07,160
ุฃุฑุจุนุฉ ุนู„ู‰ ุฃุฑุจุนุฉ sorry ุฃุฑุจุนุฉ ุนู„ู‰ ุฃุฑุจุนุฉ ููŠ ุงู„ู€ log ุงู„ู€
451
00:35:07,160 --> 00:35:13,210
binary ู„ู„ู€ ุฃุฑุจุนุฉ ุนู„ู‰ ุฃุฑุจุนุฉ ุงู„ู„ูŠ ู‡ูŠ ูˆุงุญุฏ ุตูุฑ ุฒุงุฆุฏ ุตูุฑ
452
00:35:13,210 --> 00:35:21,090
ุนู„ู‰ ุฃุฑุจุนุฉ ููŠ ุงู„ู€ log ุตูุฑ ุนู„ู‰ ุฃุฑุจุนุฉ ูˆู…ู† ุซู… ุงู„ู‚ูŠู…ุฉ ู‡ุฐู‡
453
00:35:21,090 --> 00:35:24,430
ู‡ุชุฑูˆุญ ุนู†ุฏูŠ ูˆู‡ุฐู‡ ู‡ูŠ ู†ูุณ ุงู„ู„ูŠ ููˆู‚ ุจุณ ู…ุน ุชุบูŠูŠุฑ ุงู„ู€
454
00:35:24,430 --> 00:35:30,110
terms ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ุง ุจู†ูุณ ุงู„ุญุณุจุฉ ูˆุจุงู„ุชุงู„ูŠ ุฃู†ุง
455
00:35:30,110 --> 00:35:33,450
ุญุณุจุชู‡ุง ููŠ ุงู„ุฌุฏูˆู„ ูุทู„ุนุช ู…ุนุงูŠุง ู‡ูŠู‡ุง
456
00:35:39,160 --> 00:35:43,280
ุฎู„ุตุชุŸ ู„ุฃ ู„ุณู‡ ู…ุฎู„ุตุด ุจู‚ุงู„ ุนู„ูŠ ุฎุทูˆุฉ ูˆุงุญุฏุฉ ุนุดุงู† ุฃุนุฑู
457
00:35:43,280 --> 00:35:51,340
ุฃู† ุงู„ู€ gain ุชุจุนุช ุงู„ู€ age ุฌุฏุงุด ุจุฏูŠ ุฃุฑูˆุญ ุฃู‚ูˆู„ู‡ ุงู„ู€ gain
458
00:35:51,340 --> 00:35:56,140
ุชุจุนุช ุงู„ู€ data set ุงู„ู„ูŠ ูƒู„ู‡ุง ุงู„ู„ูŠ ู‡ูŠ ุฌู…ุน ุงู„ู€ i ู„ู„ุชุณุนุฉ
459
00:35:56,140 --> 00:36:02,380
ูˆุฎู…ุณุฉ ุญุณุจู†ุงู‡ุง point ุชุณุนุฉ ุฃุฑุจุนุฉ ุตูุฑ ููŠ ุงู„ู€ slide
460
00:36:02,380 --> 00:36:03,020
ุงู„ุณุงุฏู‚ ู‡ูŠ
461
00:36:06,330 --> 00:36:10,670
ู„ู…ุง ุญุณุจู†ุงู‡ุง ู‡ู†ุง ู„ูƒู„ ุงู„ู€ data set ุงู„ู€ gain ุฃูˆ ุงู„ู€
462
00:36:10,670 --> 00:36:13,470
intro ู„ูƒู„ ุงู„ู€ data set ุญุณุจุช ุงู„ุขู† ุงู„ู€ intro ู„ู€ ุงู„ู€ age
463
00:36:13,470 --> 00:36:19,210
ู‡ูŠู‡ุง ุงู„ุขู† ุงู„ุฎุทูˆุฉ ุงู„ู„ูŠ ู‡ุดุชุบู„ ุนู„ูŠู‡ุง ุฃู†ู‡ ุจุฏูŠ ุฃุฌูŠุจ ุงู„ู€
464
00:36:19,210 --> 00:36:23,510
information gain ู„ูƒู„ ูˆุงุญุฏ ููŠู‡ู… ุทุจุนุง ุงู„ู…ูุฑูˆุถ ุฃู†ุง
465
00:36:23,510 --> 00:36:27,350
ู…ู…ูƒู† ุฃุดุชุบู„ ุงู„ุฎุทูˆุฉ ู‡ุงูŠ ุชุจุงุนุง ุฃุฌุฑู‘ุญ ุฃู‚ูˆู„ู‡ ู…ุจุงุดุฑุฉ ุงู„ุขู†
466
00:36:27,350 --> 00:36:35,550
ุงู„ู€ information gain ู„ู„ู€ age ุชุณุงูˆูŠ
467
00:36:36,360 --> 00:36:40,700
ุฃูˆ ุงู„ู€ Gain ู„ู„ู€ Age ุชุณุงูˆูŠ ุงู„ู€ Entropy ู„ูƒู„ ุงู„ู€ data
468
00:36:40,700 --> 00:36:46,120
set ู†ุงู‚ุต ุงู„ู€ Entropy ุชุจุน ุงู„ู€ Age ุงู„ู€
469
00:36:46,120 --> 00:36:53,940
Gain ู„ู„ู€ Age ุชุณุงูˆูŠ ุงู„ู€ Entropy ู„ู„ู€ data set 0.94 ู†ุงู‚ุต
470
00:36:53,940 --> 00:37:02,900
ุงู„ู€ Entropy ู„ู€ ุงู„ู€ Age ุงู„ู„ูŠ ู…ูˆุฌูˆุฏ ุนู†ุฏูŠ 0.694 ุชุณุงูˆูŠ ุทุจุนุง
471
00:37:02,900 --> 00:37:06,620
ู…ู…ูƒู† ุชุดุชุบู„ ุงู„ุฎุทูˆุฉ ุงู„ุชุงู„ูŠุฉ ุชุดุชุบู„ ู…ุน ุงู„ู€ income ุจู†ูุณ
472
00:37:06,620 --> 00:37:09,200
ุงู„ู€ concept ุงู„ู€ income ููŠ ุนู†ุฏูŠ ุจุฑุถู‡ .. ุจุฑุถู‡ ู‡ู†ุง ููŠ
473
00:37:09,200 --> 00:37:12,420
ุนู†ุฏูŠ three different values ุงู„ู€ income ููŠ ุนู†ุฏูŠ
474
00:37:12,420 --> 00:37:16,340
three different values high ูˆ medium ูˆ low
475
00:37:16,340 --> 00:37:19,840
ุชูˆุฒูŠุนุชู‡ู… ุจู†ูุณ ุงู„ูƒูŠููŠุฉ ูˆุจุญุณุจ ุงู„ู€ entropy ู„ู€ ุงู„ู€ yes ูˆ
476
00:37:19,840 --> 00:37:28,780
ุงู„ู€ no ูˆู‡ูƒุฐุง ู‡ูŠ ุญุณุจุช ุงู„ู€ entropy ู„ู€ ุงู„ู€ ..Informat ..
477
00:37:28,780 --> 00:37:32,880
ู„ู„ู€ income ุงู„ู€ intro ุจูŠ ู„ู€ ุงู„ู€ student ุงู„ู€ intro ุจูŠ ู„ู€
478
00:37:32,880 --> 00:37:39,660
ู…ูŠู† ู„ู€ ุงู„ู€ credit rating ุงู„ุขู† ุฎุทูˆุฉ ุชุงู„ูŠุฉ ู‡ุฑูˆุญ ุฃุญุณุจ
479
00:37:39,660 --> 00:37:45,500
ุงู„ู€ information gain ุฃูˆ ุงู„ู€ gain ุชุจุนุช ุงู„ู€ age ูˆ ุงู„ู€
480
00:37:45,500 --> 00:37:50,660
gain ุชุจุนุช ู‡ู†ุง ููŠ slides ู…ูู‚ูˆุฏุฉ ุงู„ู…ูุฑูˆุถ ุฃู†ุง ุดูƒู„ูŠ
481
00:37:50,660 --> 00:37:54,420
ู†ุณูŠุชู‡ุง ุฃูˆ ู…ุง ุดุงุจู‡ ู„ุฃ ู‡ูŠ ู†ูุณ ุงู„ู€ .. okay ุจุณ ู…ุด ู…ูุตู„ุฉ
482
00:37:54,420 --> 00:38:01,550
ูุญุณุจุช ุงู„ู€ gain ู„ู„ู€ age ุงู„ู€ gain ู„ู„ู€ age ู‡ูŠ
483
00:38:01,550 --> 00:38:08,250
ุนุจุงุฑุฉ ุนู† ุงู„ู€ entropy ู„ูƒู„ ุงู„ู€ data 0.694 ู†ุงู‚ุต ุงู„ู€
484
00:38:08,250 --> 00:38:13,450
entropy ุชุจุนุช ุงู„ู€ age ุงู„ู„ูŠ ู‡ูŠ 0.694 ู…ู† ุงู„ุฃู„ู ูˆู‡ูŠูƒูˆู†
485
00:38:13,450 --> 00:38:19,370
ุงู„ูุฑู‚ ุจูŠู†ู‡ู… 0.246 ู…ู† ุงู„ุฃู„ู ูˆ ุฑูˆุญุช ุญุณุจุช ุงู„ู€ income ุฃูˆ ุงู„ู€
486
00:38:19,370 --> 00:38:25,090
gain ู„ู„ู€ income ุงู„ู€ gain ู„ู„ู€ student ูˆ ุงู„ู€ gain ู„ู„ู€
487
00:38:25,090 --> 00:38:29,630
credit rating ู„ุงุญุธ ุงู„ู€ credit rating ุฃู‚ู„ ู…ุง ูŠู…ูƒู†
488
00:38:29,630 --> 00:38:36,270
ุฃู‚ู„ ุฃุตุบุฑ ูˆุงุญุฏุฉ ู…ู† ุงู„ุนู†ุงุตุฑ ุงู„ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ุฃู‚ู„
489
00:38:36,270 --> 00:38:40,830
ูˆุงุญุฏุฉ ู…ุน ุงู„ู€ income ู…ุงู„ู‡ุงุด ูƒุงู† ู„ู‡ุงุด ุฃู‚ู„ ุชุฃุซูŠุฑุง
490
00:38:42,420 --> 00:38:47,460
ุงู„ููƒุฑุฉ ุฃู† ุฃู†ุง ุจุฏูŠ ุฃุฑูˆุญ ุขุฎุฐ ุฃูˆ ุจุฏูŠ ุฃุนู…ู„ split ุนู„ู‰
491
00:38:47,460 --> 00:38:56,160
ุงู„ู€ maximum gain ู„ู„ู€ attributes ู…ูŠู†
492
00:38:56,160 --> 00:39:05,760
ุงู„ู€ maximumุŸ ู‡ูŠ ู„ุฃู† 0.246 ู…ู† 1000 ุฃูƒุจุฑ ู…ู† ุฃูƒุจุฑ ู‚ูŠู…ุฉ
493
00:39:05,760 --> 00:39:09,680
ู…ูˆุฌูˆุฏุฉ ููŠู‡ู… ุงู„ู„ูŠ ู‡ูŠ ู‡ู†ุง ู…ุน ุงู„ู€ student ูˆู‡ุฐุง ุจุชุนุทูŠู†ูŠ
494
00:39:09,680 --> 00:39:12,040
ุฅุดุงุฑุฉ ุฃู† ู…ู…ูƒู† ุงู„ู€ student ุชูƒูˆู† ู‡ูŠ ุงู„ู€ next element
495
00:39:12,040 --> 00:39:14,920
ุงู„ู„ูŠ ุฃู†ุญู…ู„ ุนู„ูŠู‡ splitting ู„ูƒู† ู…ุด ู‚ุถูŠุชูŠ ููŠ ุงู„ุขุฎุฑ
496
00:39:14,920 --> 00:39:26,420
ุฃู†ุง ู‡ุงูŠ ุงู„ุขู† ุงู„ุขู† ุงู„ู€ data set ู‡ุฑุฌุน ู„ู‡ุง ุงู„ู€
497
00:39:26,420 --> 00:39:30,660
data set ุงู„ุขู† ู‡ู†ุนู…ู„ู‡ุง partitioning ู‡ูŠ ุงู„ู€ data set
498
00:39:30,660 --> 00:39:37,180
ูƒูŠู ุงู„ู€ partition ุชุจุนุชูŠุŸ ุจูŠุจู‚ู‰ ุฅู†ู‡ ู‚ุงู„ ู„ูŠ ุฃู† ุงู„ู€ age
499
00:39:37,180 --> 00:39:43,030
ู‡ูŠ ู‡ุชูƒูˆู† ุงู„ุฃุณุงุณ ูƒู„ ุงู„ู€ rows ุงู„ู…ุญุงุทุฉ ุจุงู„ู„ูˆู† ุงู„ุฃุญู…ุฑ ู‡ุฐู‡
500
00:39:43,030 --> 00:39:49,510
ุฃูˆ ุจูŠู† ู‚ูˆุณูŠู† ุชุจุนุช ุงู„ู€ youth ู‡ุชู…ุซู„
501
00:39:49,510 --> 00:39:56,030
one data set ุฎู…ุณุฉ
502
00:39:56,030 --> 00:40:00,950
rows ุชู…ุงู…ุŸ
503
00:40:00,950 --> 00:40:05,930
ุจุนุฏ ู‡ูŠูƒ ุงู„ู€ middle age ู„ุญุงู„ู‡ู… ุงู„ู„ูŠ ุจุงู„ู„ูˆู† ุงู„ุฃุฒุฑู‚
504
00:40:05,930 --> 00:40:07,010
ุนู…ุงู„ูŠ ุจุฃุญูˆุท ุนู„ูŠู‡ู…
505
00:40:15,180 --> 00:40:19,840
ู‡ุฏูˆู„ุฉ ุฃุฑุจุนุฉ .. ุฃุฑุจุนุฉ ูˆุถู„ูˆุง ุงู„ู€ senior ุงู„ุนู†ุงุตุฑ
506
00:40:19,840 --> 00:40:23,460
ุงู„ุจุงู‚ูŠุฉ ูŠุนู†ูŠ ุจูŠู† ู‚ูˆุณูŠู† ุฃู† ุงู„ู€ data set ุจู‚ู‰ ู‡ุชู†ุฌุณู…
507
00:40:23,460 --> 00:40:28,020
ุงู„ุขู† ุจุนุฏ ู…ุง ุฃุฎุฏุช ุงู„ู€ root ุฃู†ุง ู‡ูŠู‡ุง ุจู‚ูˆู„ ุงู„ู€ age ู‡ูŠ
508
00:40:28,020 --> 00:40:32,220
ุงู„ุฃุณุงุณ ู„ุฃู† ู‡ูŠ ุตุงุญุจุฉ ุงู„ุฃูƒุจุฑ gain ู‡ูŠู‡ุง ูุจุฏูŠ ุฃุฌุณู… ุงู„ู€
509
00:40:32,220 --> 00:40:35,000
data set ุจู‚ู‰ ู„ู„ู€ three values ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ูŠุนู†ูŠ
510
00:40:35,000 --> 00:40:38,940
ุงู„ู€ youth ูˆ ุงู„ู€ middle age ูˆ ุงู„ู€ senior ู…ู…ุชุงุฒ
511
00:40:44,580 --> 00:40:48,980
ุฌุณู…ู†ุงู‡ู… ู‡ุฐู‡ ุงู„ู€ data set ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ุงู„ุขู† ุนู„ู‰
512
00:40:48,980 --> 00:40:54,480
ุงู„ุณุฑูŠุน ุดูˆ ู‡ุฑูˆุญ ุฃุณุงูˆูŠ ู‡ุงุฎุฏ ูƒู„ data set ู„ุฃู† ู„ุงุญุธ ุฃู†
513
00:40:54,480 --> 00:40:56,040
ุนู…ูˆุฏ ุงู„ู€ student ุงุฎุชูุช
514
00:40:58,670 --> 00:41:03,170
ุนู…ูˆุฏ ุงู„ู€ age ุงุฎุชูุช ..ุนู…ูˆุฏ ุงู„ู€ age ุงุฎุชูุช ..ุงู„ุขู† ูƒู„
515
00:41:03,170 --> 00:41:08,270
ูˆุงุญุฏุฉ ู…ู† ุงู„ู€ data set ู‡ุดุชุบู„ ุนู„ูŠู‡ุง ุจุดูƒู„ ู…ุณุชู‚ู„ ู„ุญุฏ ู…ุง
516
00:41:08,270 --> 00:41:13,150
ุฃุญู‚ู‚ ูˆุงุญุฏ ู…ู† ุงู„ุดุฑูˆุท ุงู„ุซู„ุงุซุฉ ุฃู…ุง ูุนู„ูŠุง ูƒู„ ุงู„ู€
517
00:41:13,150 --> 00:41:17,690
attributes ุฃูˆ ูƒู„ ุงู„ู€ symbols ุชู†ุชู…ูŠ ู„ู†ูุณ ุงู„ู€ class ุฃูˆ
518
00:41:17,690 --> 00:41:21,290
ู…ุง ุถู„ุด ููŠู‡ ุนู†ุฏูŠ more attributes ุฃู†ุง ุฌุณู…ู‡ุง ุฃูˆ ู…ุง ุถู„ุด
519
00:41:21,290 --> 00:41:24,450
ููŠู‡ ุนู†ุฏูŠ rows ุจุนุฏ ู‡ูŠูƒ ูŠุนู†ูŠ ุจูŠู† ู‚ูˆุณูŠู† ุญุงุฌุฉ ู‡ู†ุง ุงู„ุขู†
520
00:41:24,450 --> 00:41:28,890
ูˆู‡ุดุชุบู„ ุนู„ูŠู‡ุง ุจุดูƒู„ ู…ุณุชู‚ู„ ู‡ุฐู‡ ุงู„ุขู† new data set
521
00:41:28,890 --> 00:41:34,490
ู‡ุฃุญุณุจ ู„ู‡ุง information ู„ู…ูŠู†ุŸ ู‡ุฐู‡ ูƒู„ู‡ุง ุฎู…ุณ ุนู†ุงุตุฑ ุงุซู†ูŠู†
522
00:41:34,490 --> 00:41:39,330
ูˆุซู„ุงุซุฉ ู‡ุฐู‡ ุงู„ู€ information ุงู„ู„ูŠ ุฃุฏูŠู‡ ูƒู„ู‡ุง ุงู„ู€ ID
523
00:41:39,330 --> 00:41:42,630
ุชุจุนุชูŠ I
524
00:41:43,610 --> 00:41:49,410
ุซู„ุง .. ุงุซู†ูŠู† ู„ู„ู€ yes ูˆุซู„ุงุซุฉ ู„ู„ู€ no ูˆู‡ุฐุง ูŠุณุงูˆูŠ ุณุงู„ุจ ููŠ
525
00:41:49,410 --> 00:41:55,770
ู…ุฌู…ูˆุน .. ููŠ ู…ุฌู…ูˆุน ุฃูˆ ุจู„ุงุด ู†ุญุท ุงู„ู…ุฌู…ูˆุน ู‡ูŠู‡ุง ุงุซู†ูŠู†
526
00:41:55,770 --> 00:42:02,790
ุนู„ู‰ ุฎู…ุณุฉ ููŠ ุงู„ู€ log ุงุซู†ูŠู† ุนู„ู‰ ุฎู…ุณุฉ ุฒุงุฆุฏ ุซู„ุงุซุฉ ุนู„ู‰ ุฎู…ุณุฉ
527
00:42:02,790 --> 00:42:09,060
ููŠ ุงู„ู€ log ุซู„ุงุซุฉ ุนู„ู‰ ุฎู…ุณุฉ ู‡ุฐู‡ ุงู„ู…ุนู„ูˆู…ุงุช ู„ูƒู„ ุงู„ู€ data set
528
00:42:09,060 --> 00:42:12,940
ู‡ุฐู‡ as all ู…ู…ุชุงุฒ ุงู„ุขู† ูƒู… attribute ู…ูˆุฌูˆุฏ ุนู†ุฏูŠุŸ
529
00:42:12,940 --> 00:42:16,060
ุนู†ุฏูŠ three different attributes ุนู†ุฏูŠ ุงู„ู€ income ูˆ
530
00:42:16,060 --> 00:42:19,500
ุนู†ุฏูŠ ุงู„ู€ age ุนููˆุง ุงู„ู€ student ูˆ ุงู„ู€ credit rating
531
00:42:19,500 --> 00:42:23,640
ู‡ุฑูˆุญ ุฃุญุณุจ ุงู„ู€ information ู„ู…ูŠู†ุŸ ู‡ุฑูˆุญ ุฃุจุฏุฃ ุฃุจู†ูŠ
532
00:42:23,640 --> 00:42:27,500
ุงู„ุฌุฏูˆู„ ุงู„ุขู† ู„ู„ู€ attribute ุงู„ุฃูˆู„ ู…ุด ุงุชูู‚ู†ุง ู‡ูŠูƒ ู‡ุดุชุบู„
533
00:42:27,500 --> 00:42:31,360
ู…ุน ูƒู„ ุฌุฏูˆู„ ุนู„ู‰ ุงู„ุณุฑูŠุน ุฃู†ุง ู‡ุดุชุบู„ ุจุณ ู…ุน ู‡ุฐู‡ ูˆุงู„ุฎุทูˆุฉ
534
00:42:31,360 --> 00:42:34,900
ุงู„ุซุงู†ูŠุฉ ุญุงุทุทู‡ุง ู„ู„ุจุงู‚ูŠุฉ ุญุงุทุทู‡ุง ูƒูˆุงุฌุจ ุงู„ุขู†
535
00:42:37,170 --> 00:42:42,950
ู…ุน ุงู„ุนู…ูˆุฏ ุงู„ุฃูˆู„ ุงู„ู€ income ู‚ู„ุช
536
00:42:42,950 --> 00:42:49,250
ุงู„ู€ value ุชุจุนุช ุงู„ู€ income ุจุนุฏ ู‡ูŠูƒ ููŠ ุนู†ุฏูŠ ุงู„ู€ yes
537
00:42:49,250 --> 00:42:55,150
ูˆุนู†ุฏูŠ ุงู„ู€ no ูˆุนู†ุฏูŠ ุงู„ู€ intro ุจู„ุง ุงู„ู€ yes ูˆุงู„ู€ no ุจู†ุงุก
538
00:42:55,150 --> 00:42:59,090
ุนู„ู‰ ุงู„ุนุฏุฏ ุงู„ู„ูŠ ู…ูˆุฌูˆุฏ ุนู†ุฏูŠ ูˆุจุงู„ุชุงู„ูŠ ูƒู… value ู…ูˆุฌูˆุฏ
539
00:42:59,090 --> 00:43:05,010
ุนู†ุฏูŠ ุฃู†ุง ู‡ู†ุง ุนู†ุฏูŠ low ูˆ medium ูˆ high ู‡ูŠ
540
00:43:05,010 --> 00:43:05,290
low
541
00:43:08,440 --> 00:43:16,120
medium ูˆ high ุนุฏ ุงู„ุนู†ุงุตุฑ ุงู„ู…ูˆุฌูˆุฏุฉ ู…ุน ุจุนุถู†ุง ุงู„ุขู† ู…ุน
542
00:43:16,120 --> 00:43:24,500
ุงู„ู€ low ุนู†ุฏูŠ ู‚ูŠู…ุฉ ูˆุงุญุฏุฉ ูู‚ุท ู„ู…ูŠู† ุจุชู†ุชู…ูŠ ู„ู„ู€ yes ูˆุงุญุฏ
543
00:43:24,500 --> 00:43:33,560
yes ูˆู‡ู†ุง ุตูุฑ ู…ุนู†ุงุชู‡ ุงู„ู€ entropy ู„ูˆุงุญุฏ ูˆุตูุฑ ุขุฌูŠ
544
00:43:33,560 --> 00:43:35,780
ู„ู€ ุงู„ู€ medium medium ุนู†ุฏูŠ ุงุซู†ูŠู†
545
00:43:43,250 --> 00:43:50,310
ูˆุงุญุฏ yes ูˆูˆุงุญุฏ no ู…ุนู†ุงุชู‡ entropy ู„ูˆุงุญุฏ ูˆูˆุงุญุฏ ูˆุถู„
546
00:43:50,310 --> 00:43:53,950
ููŠ ุนู†ุฏูŠ high ุงุซู†ูŠู† ูˆุจูŠู†ุชู…ูˆุง ู„ู†ูุณ ุงู„ู€ class
547
00:43:53,950 --> 00:43:59,710
ู…ุนู†ุงุชู‡ ุตูุฑ ูˆุงุซู†ูŠู† entropy ู„ุตูุฑ ูˆุงุซู†ูŠู† ูˆู‡ุฐุง ุจุฐูƒุฑ ุฃู†
548
00:43:59,710 --> 00:44:03,650
ู‡ุฐุง ุงู„ู‚ุงู†ูˆู† ุงู„ุณุงุจู‚ ุงู„ู„ูŠ ุงุนุชู…ุฏุช ุนู„ูŠู‡ ุญุตู„ ุจุญุณุจ ุงู„ู€
549
00:44:03,650 --> 00:44:08,810
gain ุงู„ุขู† ุฃูˆ ุจุญุณุจ ุงู„ู€ information ู„ู„ู€ attribute ุงู„ู„ูŠ
550
00:44:08,810 --> 00:44:15,860
ู…ูˆุฌูˆุฏ ุนู†ุฏูŠ ู…ู† ุฎู„ุงู„ ุงู„ู€ summation ูƒู…ุงู† ู…ุฑุฉ ุนุฏุฏ ุงู„ู€
551
00:44:15,860 --> 00:44:21,840
data set ุฒูŠ ู…ุง ู‚ู„ู†ุง ุฌุฏุงุด ุฌู…ุนุฉ ุงู„ุฎูŠุฑ ุฎู…ุณุฉ ุงู„ุขู†
552
00:44:21,840 --> 00:44:31,420
ูˆุงุญุฏ ุนู„ู‰ ุฎู…ุณุฉ ููŠ ุงู„ู€ I ูˆุงุญุฏ ูˆุตูุฑ ุฒุงุฆุฏ ู‡ุฐู‡ ุงู„ู€
553
00:44:31,420 --> 00:44:40,050
information ุงุซู†ูŠู† ุนู„ู‰ ุฎู…ุณุฉ ููŠ ุงู„ู€ I ูˆุงุญุฏ ูˆูˆุงุญุฏ ุฒุงุฆุฏ
554
00:44:40,050 --> 00:44:49,530
ุงุซู†ูŠู† ุนู„ู‰ ุฎู…ุณุฉ ุงุซู†ูŠู† ุนู„ู‰ ุฎู…ุณุฉ ุตุญูŠุญ ุงุซู†ูŠู† ุนู„ู‰ ุฎู…ุณุฉ
555
00:44:49,530 --> 00:44:58,530
ููŠ ุงู„ู€ I ุตูุฑ
556
00:44:58,530 --> 00:45:04,690
ูˆุงุซู†ูŠู† ุจุญุตู„ ุนู„ู‰ ุงู„ู€ information gain ุงู„ู„ูŠ ู…ูˆุฌูˆุฏ
557
00:45:04,690 --> 00:45:12,160
ุนู†ุฏูŠ ู‡ู†ุง ุจุนุฏ ู‡ูŠูƒ ุจู‚ูˆู„ู‡ ุงู„ู€ gain ุชุจุนุช ุงู„ู€ income ู‡ุชู…ุซู„
558
00:45:12,160 --> 00:45:21,900
ุงู„ู€ I ุงู„ู„ูŠ ุนู†ุฏูŠ ููˆู‚ ู†ุงู‚ุต ุงู„ู€ I ู„ู„ู€ income ุงู„ู„ูŠ
559
00:45:21,900 --> 00:45:28,380
ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู‡ุง ู‡ุฃุญุณุจู‡ุง ูˆุจุงู„ุชุงู„ูŠ ุจุฃุญุณุจ ู„ู‡ุฐู‡ ูˆุจุฃุญุณุจ ู„ู„ู€
560
00:45:28,380 --> 00:45:31,780
ุงู„ุนู†ุงุตุฑ
561
00:45:31,780 --> 00:45:37,640
ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ุชู…ุงู…ุง ู„ู„ู€ student ุจุนูŠุฏ ุงู„ูƒุฑุฉ ูˆ ู„ู„ู€
562
00:45:37,640 --> 00:45:41,780
credit rating ูˆุตุงุญุจ ุงู„ู€ attribute ุตุงุญุจ ุฃูƒุจุฑ gain
563
00:45:41,780 --> 00:45:46,480
ู‡ูˆ ุงู„ู„ูŠ ุญูŠูƒูˆู† ูุนู„ูŠุง ุฃู†ุง ู‡ุฃุนุชู…ุฏ ูˆูŠู† ููŠ ุงู„ู€ .. ููŠ
564
00:45:46,480 --> 00:45:49,740
ุงู„ุฑุณู…ุฉ ุฃูˆ ููŠ ุงู„ู€ decision node ุงู„ุชุงู„ูŠุฉ ุญุณุจ ุงู„ุญุณุจุฉ
565
00:45:49,740 --> 00:45:54,200
ุชุจุนุชูŠ ุญุณุจู†ุงู‡ุง ุณุงุจู‚ุง ู„ุงุฒู… ุฃู†ุชู… ุชูƒู…ู„ูˆู‡ุง ู„ู„ุงุฎุฑ ุญุณุจ
566
00:45:54,200 --> 00:45:57,740
ุงู„ุญุณุจุฉ ุชุจุนุชูŠ ุงู„ู€ student ุญุตู„ุช ุฃุนู„ู‰ gain ุงู„ู€ student
567
00:45:57,740 --> 00:46:02,680
ุชุจุนุชูŠ ุญุตู„ุช ุฃุนู„ู‰ gain ูˆุจุงู„ุชุงู„ูŠ ุฃู†ุง ุงู„ุขู† ู‡ู†ุง ู‡ุฃุตูŠุฑ
568
00:46:02,680 --> 00:46:07,260
ููŠ ุนู†ุฏูŠ ุงู„ู€ student ู‡ูŠ ุงู„ู€ Internal node ุงู„ุฌุงูŠุฉ ูˆููŠู‡ุง
569
00:46:07,260 --> 00:46:13,680
two different values ู‡ุฃู‚ุณู… ุงู„ู€ data set ุจุนุฏ ู‡ูŠูƒ ู‡ุฃู‚ุณู…
570
00:46:13,680 --> 00:46:16,860
ุงู„ู€ data set ุชุจุนุง ู„ู„ู€ nodes ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ู‡ูŠูƒ ูˆ
571
00:46:16,860 --> 00:46:20,960
ุจู‡ูŠูƒ ุตุงุฑุช ูƒู„ ุงู„ู€ nodes ุชู†ุชู…ูŠ ู„ู†ูุณ ุงู„ู€ class ูƒู„ ุงู„ู€
572
00:46:20,960 --> 00:46:24,620
samples ุชู†ุชู…ูŠ ู„ู†ูุณ ุงู„ู€ class ูู‡ู†ุง ุจูˆู‚ู ู‡ุฐู‡ already
573
00:46:24,620 --> 00:46:28,260
ูƒู„ู‡ุง ุชู†ุชู…ูŠ ู„ู†ูุณ ุงู„ู€ class ูุฃู†ุง ูˆู‚ูุช ู‡ู†ุง ูˆูˆู‚ูุช ู‡ู†ุง
574
00:46:28,260 --> 00:46:31,540
ู‡ุชูƒูˆู† ุงู„ู€ final tree ุชุจุนุชูŠ ุทุจุนุง ู‡ูŠ ุงู„ุญุณุจุฉ ุงู„ู„ูŠ
575
00:46:31,540 --> 00:46:36,430
ุนู†ุฏู†ุงู‡ุง ู…ุฑุฉ ุซุงู†ูŠุฉ ุฃู†ุง ููŠ ุงู„ุขุฎุฑ ุงู„ู€ Tree ุชุจุนุชูŠ ุฃุญุตู„
576
00:46:36,430 --> 00:46:40,330
ุนู„ูŠู‡ุง ุงู„ู„ูŠ ุฅุญู†ุง ุดูู†ุงู‡ุง ู…ุณุจู‚ุง ุงู„ู…ูุฑูˆุถ ู‚ุจู„ ู…ุง ูŠุจุฏุฃ
577
00:46:40,330 --> 00:46:50,130
ุจุงู„ุดุบู„ ุงู„ู„ูŠ ู‡ูŠ ู‡ุฐู‡ ุงู„ู€ income ู…ุง ุจูŠู†ุชุด ุนู†ุฏูŠ ู„ุฃู†
578
00:46:50,130 --> 00:46:54,770
ูุนู„ูŠุง ูˆุฒู†ู‡ุง ูƒุงู† ู„ุง ูŠุฐูƒุฑ ู…ู‚ุงุฑู†ุฉ ุจุงู„ู€ data set ูˆู„ู…ุง
579
00:46:54,770 --> 00:46:57,570
ุฃู†ุง ู…ุง ุถู„ุด ุนู†ุฏูŠ rows ุฃูˆ ู…ุง ุถู„ุด ุนู†ุฏูŠ sample ุฃุฑูˆุญ
580
00:46:57,570 --> 00:47:01,650
ุฃุฌุณู…ู‡ุง ุงู„ุขู† ุนุดุงู†
581
00:47:02,490 --> 00:47:06,390
ู…ุง ู†ุทูˆู„ุด ุนู„ูŠูƒู… ุจุนุฏ ู…ุง ุญุณุจู†ุง ุงู„ุนู†ุงุตุฑ ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ
582
00:47:06,390 --> 00:47:11,430
ุนู†ุฏู‡ุง ููŠ ุงู„ู€ continuous attributes ููŠ ุญุงู„ุฉ ุงู„ู€
583
00:47:11,430 --> 00:47:14,470
attributes ุงู„ู„ูŠ ุนู†ุฏูƒ continuous attribute ุฅูŠุด ุงู„ุญู„ุŸ
584
00:47:14,470 --> 00:47:19,610
ุงุนู…ู„ discretization ุงุนู…ู„ู‡ุง categories ุฌุณู…ู‡ุง ู„ูุฆุงุช
585
00:47:19,610 --> 00:47:22,830
ุงุณุชุฎุฏู… ุงู„ู€ binning ูˆ ุฃุนุทูŠ label ู„ูƒู„ bin ูˆ ุงุดุชุบู„
586
00:47:22,830 --> 00:47:28,710
ุนู„ูŠู‡ุง ุจุชูƒุงุดูŠ ุงุดุชุบู„ ุนู„ูŠูƒ ููŠ ุญู„ ุจุณูŠุท ุฌุฏุง ุงู„ุญู„ ุจูŠู‚ูˆู„ูƒ
587
00:47:28,710 --> 00:47:33,760
ุฑุชุจ ุงู„ู€ data set ุชุจุนุง ู„ู„ู€ items ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูƒ ุฑุชุจ
588
00:47:33,760 --> 00:47:38,260
ุงู„ู€ items ุชู…ุงู… ูุตุงุฑุช ุงู„ู€ data sorted ุฅู† ููŠ ุญุงู„ ูƒุงู†ุช
589
00:47:38,260 --> 00:47:43,260
ุนู†ุฏูŠ ุงู„ู€ age ุนุจุงุฑุฉ ุนู† number ุตุงุฑ ุนู†ุฏูŠ ุฎู…ุณุฉ ูˆุนุดุฑูŠู† ูˆ
590
00:47:43,260 --> 00:47:48,780
ุซู„ุงุซูŠู† ุณุจุนุฉ ูˆุซู„ุงุซูŠู† ุฃุฑุจุนูŠู† ุงู„ุขู† ุฃู†ุช ุจู‚ู‰
591
00:47:48,780 --> 00:47:55,500
ุชูŠุฌูŠ ุชูุญุต ุจูŠู† ูƒู„ two nodes ูŠุนู†ูŠ ู‡ุชุฃุฎุฐ
592
00:47:55,500 --> 00:48:00,200
ุงู„ู€ midpoint ุงู„ู„ูŠ ุจูŠู† ู‡ุฏูˆู„ ุงู„ุงุซู†ูŠู† ุฃูˆ ุงู„ุฃุณู‡ู„ ู„ูƒ
593
00:48:01,200 --> 00:48:03,600
ูุนู„ุง ู‡ูŠ ุนุจุงุฑุฉ ุนู† Discretization ู„ูƒู†ู‡ุง Binary
594
00:48:03,600 --> 00:48:07,780
Discretization ู…ุน ุงู„ุฃุฑู‚ุงู… ุฃู†ุช ุฅูŠุด ุงู„ู…ู‚ุงุฑู†ุงุช ุชุจุนุชูƒุŸ
595
00:48:07,780 --> 00:48:12,460
ุฃู…ุง ู‡ุชู‚ูˆู„ ู„ูŠ ุฃู‚ู„ ุฅุฐุง ู‚ู„ุช ุฃู‚ู„ ู…ู† ูƒุฏู‡ ูู‡ูŠ ุฃูƒุจุฑ ุฃูˆ
596
00:48:12,460 --> 00:48:16,340
ุชุณุงูˆูŠ ูƒุฏู‡ ุฅุฐุง ู‚ู„ุช ุฃูƒุจุฑ ู…ู† ุฃูƒุจุฑ ู…ู† ุฃูˆ ุชุณุงูˆูŠ ูƒุฏู‡
597
00:48:16,340 --> 00:48:19,280
ูู‡ูŠ ุฃู‚ู„ ู…ู† ูƒุฏู‡ ุนูƒุณู‡ุง ุชู…ุงู…ุง ูุญูŠูƒูˆู† ุจุดุบู„ ุนู„ูŠู‡ุง ู…ุน
598
00:48:19,280 --> 00:48:23,080
binary ูŠุนู†ูŠ ุจูŠู† ู‚ูˆุณูŠู† ุจุตูŠุฑ ุจุฃุฎุฐ decision ู‡ู†ุง ุจู‚ูˆู„ู‡
599
00:48:23,080 --> 00:48:30,500
ุฃู‚ู„ ุฃูˆ ุชุณุงูˆูŠ ุฎู…ุณุฉ ูˆุนุดุฑูŠู† ุทุจ ู…ุง ู‡ูŠ ุฎู…ุณุฉ ูˆุนุดุฑูŠู† ุจุงุฌูŠ ู‡ู†ุง
600
00:48:33,210 --> 00:48:39,230
ุฃู‚ู„ ุฃูˆ ุชุณุงูˆูŠ ุซู„ุงุซูŠู† ู…ุนุชุจุฑ ู‡ุฏูˆู„ ููŠ partitions ูˆ
601
00:48:39,230 --> 00:48:42,190
ู‡ุฏูˆู„ ููŠ partitions ู…ุน ุงู„ู€ continuous attributes ุฅุฐุง
602
00:48:42,190 --> 00:48:45,830
ุฃู†ุช ุจุฏูƒ ุชุดุชุบู„ ู…ุน ุงู„ู€ continuous values ู…ุนู†ุงุชู‡ ุฃู†ุช
603
00:48:45,830 --> 00:48:49,990
ู‡ูŠูƒูˆู† ููŠ ุนู†ุฏูƒ too many partitions ู„ุญุฏ ู…ุง ุชุตู„ ู„ู€
604
00:48:49,990 --> 00:48:55,300
best point ุงู„ู„ูŠ ุจุชุนู…ู„ split ูŠุนู†ูŠ ู‡ุฃุญุณุจ ุงู„ุขู† ุงู„ุฌุณู…
605
00:48:55,300 --> 00:48:58,920
ูƒุฏู‡ ูƒุงู†ุช ู‡ู†ุง ูˆู„ุง ู„ู…ุง ูƒุงู†ูˆุง ุงุซู†ูŠู† ูˆุซู„ุงุซุฉ ูˆู„ู…ุง
606
00:48:58,920 --> 00:49:03,160
ูƒุงู†ูˆุง ุซู„ุงุซุฉ ูˆุงุซู†ูŠู† ูˆุฃุฌุฑุจ ูƒู„ู‡ู… ูˆุขุฎุฐ ุฃุนู„ู‰ gain
607
00:49:03,160 --> 00:49:06,900
ููŠู‡ู… ู„ุฃู† ููŠ ุงู„ุขุฎุฑ ุฃู†ุง ุจุฏูˆุฑ ุนู„ู‰ ุงู„ู€ gain ู„ูƒู„ ุงู„ู€ data
608
00:49:06,900 --> 00:49:12,320
set ุชุจุนุชูŠ ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏูŠ ู‡ู†ุง ุงู„ุขู† ู‡ุฏู ุงู„ู…ูˆุถูˆุน ุงู„ู€
609
00:49:12,320 --> 00:49:15,810
splitting ู„ู„ู€ continuous values ู„ูƒู† ุงู„ู€ Information
610
00:49:15,810 --> 00:49:21,330
Gain ุฏุงุฆู…ุง ุจูŠุญุงุฒ ู„ู„ู€ attributes ุฃูˆ ู„ู„ู€ test ุงู„ู„ูŠ
611
00:49:21,330 --> 00:49:25,230
ุจูŠูƒูˆู† ููŠู‡ุง two ุฃูˆ ููŠู‡ุง many outcomes ุงู„ู„ูŠ ููŠู‡ุง
612
00:49:25,230 --> 00:49:30,490
values ูƒุซูŠุฑุฉ ุนุดุงู† ู‡ูŠ ูƒุงู†ุช ููŠ ุงู„ุฃูˆู„ ุนู†ุฏูŠ ุงู„ู€ age
613
00:49:30,490 --> 00:49:33,190
ูƒุงู†ุช ู‡ูŠ ุฃูƒุซุฑ ุงู„ู€ values ุงู„ู„ูŠ ู…ูˆุฌูˆุฏุฉ ุฌุงู„ูƒ ู…ู…ูƒู† ุฅุญู†ุง
614
00:49:33,190 --> 00:49:37,070
ู†ุญูŽู„ ู‡ุฐู‡ ุงู„ู…ุดูƒู„ุฉ ูˆู†ุนุชู…ุฏ ุฃูˆ ู†ุญุงูˆู„ ู†ู‚ุถูŠ ุนู„ู‰ ู…ูˆุถูˆุน
615
00:49:37,070 --> 00:49:39,930
ุฅู†ุญูŠุงุฒ ุงู„ู€ values ุงู„ูƒุซูŠุฑุฉ ุงู„ู„ูŠ ู‡ูˆ ู…ูˆุถูˆุน ุงู„ู€ gain
616
00:49:39,930 --> 00:49:43,500
ratio ุงู„ู€ Gain Ratio ููƒุฑุชู‡ุง ูŠุง ุฌู…ุงุนุฉ ุงู„ุฎูŠุฑ ุฃู†ู‡ ุฃู†ุง
617
00:49:43,500 --> 00:49:48,240
ุจุฏูŠ ุฃุฑูˆุญ ุฃุญุณุจ ุงู„ู€ split info ุฅุญู†ุง ุณุงุจู‚ุง ูƒุงู†ุช ู‡ุฐู‡
618
00:49:48,240 --> 00:49:52,240
ุงู„ู‚ูŠู…ุฉ ู…ูˆุฌูˆุฏุฉ ุนู†ุฏู†ุง ู‡ูŠ ุนุฏุฏ ุนู†ุงุตุฑ ุงู„ู€ partition ุนู„ู‰
619
00:49:52
667
00:53:11,450 --> 00:53:15,430
ู…ูˆุฌูˆุฏ ุนู†ุฏูŠ ุขุฎุฑ ุฎุทูˆุฉ ุฃูˆ ุขุฎุฑ slide ููŠ ุงู„ู…ูˆุถูˆุน
668
00:53:15,430 --> 00:53:18,630
ูƒูŠู ุจุฏูŠ ุฃุณุชุฏุนูŠู‡ุงุŒ ุงู„ุฎุทูˆุงุช ุงู„ุณุงุจู‚ุฉ ููŠ ุงู„ู€ python ู†ูุณู‡ุงุŸ
669
00:53:18,630 --> 00:53:22,590
from sklearn.tree import DecisionTreeClassifier
670
00:53:24,140 --> 00:53:27,880
ุงู„ู†ู…ูˆุฐุฌ ุงู„ู€ Decision Tree Classifier ุนู…ู„ุช ู„ู‡ fit
671
00:53:27,880 --> 00:53:31,720
ู‚ู„ุช ู„ู‡ุŒ ูˆู‡ุฐู‡ ุงู„ุฌุฒุฆูŠุฉ ููŠ ุงู„ุณุทุฑ ู‡ุฐุงุŒ ุจุชุจู†ูŠ ุงู„ุดุฌุฑุฉ.
672
00:53:31,720 --> 00:53:36,500
ุงู„ุขู† ุงู„ู€
673
00:53:36,500 --> 00:53:39,340
sample test ู†ูุณู‡ุงุŒ ู„ุฃู†ู‡ุง ุดุบุงู„ุฉุŒ ู†ูุณ ุงู„ุดูŠุกุŒ ุจุฏู†ุง ู†ุฌุฑุจู‡ุง ู…ุนุงูƒู… ู…ู†
674
00:53:39,340 --> 00:53:44,060
ุงู„ุจุฏุงูŠุฉ. ุฌุฑุจุชู‡ุง ู…ุน ุงู„ู€ kenia sniperุŒ ูˆุฌุฑุจุชู‡ุง ู…ุน ุงู„ู€ naive
675
00:53:44,060 --> 00:53:48,840
ุจุงูŠุฒ. ุจุฑูˆุญ ุชุนู…ู„ ุงู„ู€ testุŒ ูˆู‡ู†ุญุฏุฏ setosa ุจูƒู„
676
00:53:48,840 --> 00:53:53,600
ุชุฃูƒูŠุฏุŒ ูˆุจู‡ูŠูƒ ุจู†ูƒูˆู† ุฅุญู†ุง ูุนู„ูŠุงู‹ ุงู†ุชู‡ูŠู†ุง ู…ู† ู…ูˆุถูˆุน ุงู„ู€
677
00:53:53,600 --> 00:53:57,040
Decision Tree ู„ู…ุญุงุถุฑุชู†ุง ุงู„ูŠูˆู…. ุงู„ู…ุทู„ูˆุจ ู…ู†ูƒู…
678
00:53:57,040 --> 00:53:59,500
ุชุฌุฑุจูˆุง ูŠุง ุฌู…ุงุนุฉ ุงู„ุฎูŠุฑุŒ ููŠ ุนู†ุฏู†ุง different data set
679
00:53:59,500 --> 00:54:02,780
ู…ูˆุฌูˆุฏุฉ ููŠ ุงู„ู€ slide ุณุงุจู‚ุงู‹. ุฌุฑุจูˆุง ุงู„ูƒู„ุงู… ู‡ุฐุง ุนู„ูŠู‡ุงุŒ ูˆ
680
00:54:02,780 --> 00:54:05,000
ุฌุฑุจูˆุง ุงู„ูƒู„ุงู… ู‡ุฐุง ุนู„ูŠู‡ุงุŒ ูŠุนู†ูŠ ุจูŠู†ุชุฌ ุณูŠู†ุงุฑูŠูˆ
681
00:54:05,000 --> 00:54:08,470
ู‡ุฐุง ุงู„ูƒู„ุงู… ู…ุด ู‡ุชุชู‚ู†ูˆู‡ ู…ู† ู…ุฑุฉ ูˆุงุซู†ุชูŠู† ูˆุซู„ุงุซุŒ ุงู„ุดุบู„
682
00:54:08,470 --> 00:54:12,130
ุงู„ุชุงู†ูŠ ุงู„ู„ูŠ ุจุฏูŠ ุฅูŠุงู‡ ู…ู†ูƒู…ุŒ ุจุนุฏ ุชุฌุฑูŠุจ ุงู„ุนู…ู„ ูŠุจุฏูˆ
683
00:54:12,130 --> 00:54:15,730
ุชุนุชุจุฑูˆู‡ุง ูƒู€ assignment ุนู„ูŠูƒู…. ุงู„ุขู†ุŒ ูŠุจุฏูˆ ุชุฑูˆุญ ุชููƒุฑูˆุง
684
00:54:15,730 --> 00:54:18,570
ุฃูˆ ุชุฏูˆุฑูˆุง ู„ูŠ ูƒูŠู ู…ู…ูƒู† ุฃู†ุงุŒ ุฅุฐุง ูƒุงู†ุช ู‡ุฐู‡ ุนุจุงุฑุฉ ุนู† ุงู„ู€
685
00:54:18,570 --> 00:54:21,930
treeุŒ ุจุนุฏ ู…ุง ุฃู†ุง ุนู…ู„ุช ู„ู‡ุง fitุŒ ู‡ู„ ููŠ ู…ุฌุงู„ ุฃุฑุณู… ุงู„ู€ tree
686
00:54:21,930 --> 00:54:25,690
ุจุชุงุนุชู‡ุง ุจุงู„ุจุงูŠุซูˆู†ุŸ ุงู‡ุŒ ููŠ ู…ุฌุงู„ุŒ ูˆู‡ุฐู‡ ู…ุชุฑูˆูƒุฉ ู„ูƒู…. ูˆุงู„ุณู„ุงู…
687
00:54:25,690 --> 00:54:27,470
ุนู„ูŠูƒู… ูˆุฑุญู…ุฉ ุงู„ู„ู‡ ูˆุจุฑูƒุงุชู‡