add macro_index_mse

README.md

@@ -28,7 +28,7 @@ At minium, the metric requires the original time-series and the generated time-s
 >>> metric = evaluate.load("bowdbeg/matching_series")
 >>> results = metric.compute(references=references, predictions=predictions, batch_size=1000)
 >>> print(results)
-{'matching_mse': 0.15250070138019745, 'harmonic_mean': 0.15246672297315564, 'covered_mse': 0.15243275970407652, 'index_mse': 0.16772539808686357, 'matching_mse_features': [0.11976368411913872, 0.1238622735860897, 0.1235259257706047, 0.12385236248438022, 0.12241466736218365, 0.12328439290438079, 0.1232240061707885, 0.12342319803028035, 0.12235222572924524, 0.12437865819262514], 'harmonic_mean_features': [0.12010478503934609, 0.12379899085819131, 0.12321441761307182, 0.12273884163905005, 0.12256126537300535, 0.12323289686030311, 0.12323847434641247, 0.12333469339243568, 0.12273530480438972, 0.12390254295493403], 'covered_mse_features': [0.12044783449951382, 0.1237357727610885, 0.12290447662839017, 0.12164516506865233, 0.12270821492248948, 0.12318144381818667, 0.12325294591995689, 0.12324631559392285, 0.12312079021887229, 0.12343005890751833], 'index_mse_features': [0.16331894487549958, 0.1679797859239729, 0.16904075114728268, 0.16962427920551068, 0.16915910655024802, 0.16686197230602684, 0.17056311327206022, 0.1638796919248867, 0.16736730842643857, 0.16945902723670975], 'macro_matching_mse': 0.1230081394349717, 'macro_covered_mse': 0.12276730183385913, 'macro_harmonic_mean': 0.12288622128811397}
+{'matching_mse': 0.15873331613053895, 'harmonic_mean': 0.15623569099681772, 'covered_mse': 0.15381544718035087, 'index_mse': 0.16636189201532087, 'matching_mse_features': [0.13739837269222452, 0.1395309409295018, 0.13677679887355126, 0.14408421162706211, 0.1430115910456261, 0.13726657544044085, 0.14274372684301717, 0.13504614539190338, 0.13853582796877975, 0.14482307626368343], 'harmonic_mean_features': [0.1309991815519093, 0.13157175020534279, 0.12735134531950718, 0.1327483317911355, 0.1336402851605765, 0.12878380179856022, 0.1344831997941457, 0.12782689483798823, 0.12909420446395195, 0.13417435670997752], 'covered_mse_features': [0.12516953618356524, 0.12447158260731798, 0.11914118322950448, 0.12306606276504639, 0.1254216201001874, 0.12128844181049621, 0.12712643943219143, 0.12134032531607968, 0.12085741660832867, 0.12498436126166071], 'index_mse_features': [0.16968036010688156, 0.1624888691672768, 0.15926142198600082, 0.17250634507748022, 0.16713668302081525, 0.16663213728264645, 0.1596766027744231, 0.16251306560725656, 0.17160303243460656, 0.17212040269582168], 'macro_matching_mse': 0.13992172670757905, 'macro_covered_mse': 0.12328669693143782, 'macro_harmonic_mean': 0.13106733516330948, 'macro_index_mse': 0.1663618920153209}
 ```
 
 ### Inputs
@@ -51,6 +51,7 @@ Let prediction instances be $P = \{p_1, p_2, \ldots, p_n\}$ and reference instan
 - **macro_matching_mse**: (float): Average of the matching_mse_features.
 - **macro_covered_mse**: (float): Average of the covered_mse_features.
 - **macro_harmonic_mean**: (float): Average of the harmonic_mean_features.
+- **macro_index_mse**: (float): Average of the index_mse_features.
 
 #### Values from Popular Papers
 <!-- *Give examples, preferrably with links to leaderboards or publications, to papers that have reported this metric, along with the values they have reported.* -->

matching_series.py

@@ -204,6 +204,7 @@ class matching_series(evaluate.Metric):
         macro_matching_mse = statistics.mean(matching_mse_features)
         macro_covered_mse = statistics.mean(covered_mse_features)
         macro_harmonic_mean = statistics.mean(harmonic_mean_features)
+        macro_index_mse = statistics.mean(index_mse_features)
         return {
             "matching_mse": matching_mse,
             "harmonic_mean": harmonic_mean,
@@ -216,4 +217,5 @@ class matching_series(evaluate.Metric):
             "macro_matching_mse": macro_matching_mse,
             "macro_covered_mse": macro_covered_mse,
             "macro_harmonic_mean": macro_harmonic_mean,
+            "macro_index_mse": macro_index_mse,
         }