Machine learning models/Production/Wikidata item topic

Model card
This page is an on-wiki machine learning model card.
A model card is a document about a machine learning model that seeks to answer basic questions about the model.
Model Information Hub
Model creator(s)	Aaron Halfaker (User:EpochFail) and Amir Sarabadani
Model owner(s)	WMF Machine Learning Team (ml@wikimediafoundation.org)
Model interface	Ores homepage
Code	drafttopic Github, ORES training data, and ORES model binaries
Uses PII	No
In production?	Yes
Which projects?	Wikidata
This model uses item features to predict the likelihood that the item belongs to a set of topics.

Motivation

How can we predict what general topic an item is in? Answering this question is useful for various analyses of Wikidata dynamics. However, it is difficult to group a very diverse range of Wikidata items into coherent, consistent topics manually.

This model, part of the ORES suite of models, analyzes an item to predict its likelihood of belonging to a set of topics. Similar models (though not necessarily with the same performance level or topics, are deployed across about a dozen other projects. There is also a language agnostic article topic model.

This model may be useful for high-level analyses of Wikidata dynamics (pageviews, item quality, edit trends) and filtering items.

Users and uses

Use this model for

• high-level analyses of Wikidata dynamics such as pageview, item quality, or edit trends — e.g. How are pageview dynamics different between the physics and biology categories?
• filtering to relevant items — e.g. filter items only to those in the music category.

Don't use this model for

• definitively establishing what topic an items pertains to
• automated editing of items or topics without a human in the loop

Current uses

This model is a part of ORES, and generally accessible via API. It is used for high-level analysis of Wikidata, platform research, and other on-wiki tasks.

Example API call:

https://ores.wikimedia.org/v3/scores/wikidatawiki/1907686315/itemtopic

Ethical considerations, caveats, and recommendations

• This model was trained on data that is now several years old (from mid-2020). Underlying data drift may skew model outputs.
• This model uses word2vec as a training feature. Word2vec, like other natural language embeddings, encodes the linguistic biases of underlying datasets — along the lines of gender, race, ethnicity, religion etc. Since Wikidata has known biases in its text, this model may encode and at times reproduce those biases.
• This model has highly variable performance across different topics — consult the test statistics below to get a sense of inter-topic performance.

Model

Performance

Test data confusion matrix:

Test data confusion matrix
Label n True positive False positive False negative True Negative Culture.Biography.Biography* 16670 15762 908 464 46810 Culture.Biography.Women 4110 3125 985 679 59155 Culture.Food and drink 1318 613 705 126 62500 Culture.Internet culture 2966 1948 1018 140 60838 Culture.Linguistics 1466 934 532 56 62422 Culture.Literature 5367 3996 1371 404 58173 Culture.Media.Books 1974 1560 414 136 61834 Culture.Media.Entertainment 1733 857 876 162 62049 Culture.Media.Films 2295 1896 399 122 61527 Culture.Media.Media* 14383 11572 2811 1135 48426 Culture.Media.Music 2583 2027 556 247 61114 Culture.Media.Radio 1156 857 299 44 62744 Culture.Media.Software 1750 685 1065 307 61887 Culture.Media.Television 2230 1510 720 176 61538 Culture.Media.Video games 2147 1758 389 54 61743 Culture.Performing arts 1334 741 593 116 62494 Culture.Philosophy and religion 2702 1074 1628 285 60957 Culture.Sports 5925 5186 739 249 57770 Culture.Visual arts.Architecture 2648 1867 781 230 61066 Culture.Visual arts.Comics and Anime 1508 1007 501 140 62296 Culture.Visual arts.Fashion 1199 669 530 98 62647 Culture.Visual arts.Visual arts* 6070 4131 1939 554 57320 Geography.Geographical 3464 2226 1238 359 60121 Geography.Regions.Africa.Africa* 6449 4664 1785 414 57081 Geography.Regions.Africa.Central Africa 1145 697 448 83 62716 Geography.Regions.Africa.Eastern Africa 1114 704 410 56 62774 Geography.Regions.Africa.Northern Africa 1280 774 506 108 62556 Geography.Regions.Africa.Southern Africa 1244 859 385 81 62619 Geography.Regions.Africa.Western Africa 1142 774 368 75 62727 Geography.Regions.Americas.Central America 1331 707 624 87 62526 Geography.Regions.Americas.North America 7625 5064 2561 1169 55150 Geography.Regions.Americas.South America 1532 1082 450 142 62270 Geography.Regions.Asia.Asia* 11647 8432 3215 835 51462 Geography.Regions.Asia.Central Asia 1086 671 415 70 62788 Geography.Regions.Asia.East Asia 2717 1727 990 241 60986 Geography.Regions.Asia.North Asia 2076 1336 740 163 61705 Geography.Regions.Asia.South Asia 2366 1612 754 135 61443 Geography.Regions.Asia.Southeast Asia 1721 1059 662 119 62104 Geography.Regions.Asia.West Asia 2160 1473 687 129 61655 Geography.Regions.Europe.Eastern Europe 3533 2472 1061 234 60177 Geography.Regions.Europe.Europe* 12939 9372 3567 1810 49195 Geography.Regions.Europe.Northern Europe 4221 2571 1650 601 59122 Geography.Regions.Europe.Southern Europe 2438 1565 873 268 61238 Geography.Regions.Europe.Western Europe 3076 1934 1142 417 60451 Geography.Regions.Oceania 2638 1859 779 138 61168 History and Society.Business and economics 3502 1544 1958 569 59873 History and Society.Education 2243 1113 1130 255 61446 History and Society.History 3172 1154 2018 360 60412 History and Society.Military and warfare 3238 1677 1561 296 60410 History and Society.Politics and government 4590 2406 2184 329 59025 History and Society.Society 2971 897 2074 166 60807 History and Society.Transportation 3629 2615 1014 169 60146 STEM.Biology 2916 2237 679 91 60937 STEM.Chemistry 1270 690 580 138 62536 STEM.Computing 1968 828 1140 332 61644 STEM.Earth and environment 1627 918 709 114 62203 STEM.Engineering 2195 1284 911 141 61608 STEM.Libraries & Information 1174 605 569 87 62683 STEM.Mathematics 1137 307 830 107 62700 STEM.Medicine & Health 1726 769 957 180 62038 STEM.Physics 1219 448 771 107 62618 STEM.STEM* 16449 12609 3840 2766 44729 STEM.Space 1365 932 433 47 62532 STEM.Technology 3648 1396 2252 424 59872

Test data sample rates:

Test data sample rates
Label Sample Population Culture.Biography.Biography* 0.261 0.12 Culture.Biography.Women 0.064 0.015 Culture.Food and drink 0.021 0.003 Culture.Internet culture 0.046 0.004 Culture.Linguistics 0.023 0.008 Culture.Literature 0.084 0.015 Culture.Media.Books 0.031 0.004 Culture.Media.Entertainment 0.027 0.004 Culture.Media.Films 0.036 0.012 Culture.Media.Media* 0.225 0.055 Culture.Media.Music 0.04 0.021 Culture.Media.Radio 0.018 0.002 Culture.Media.Software 0.027 0.001 Culture.Media.Television 0.035 0.009 Culture.Media.Video games 0.034 0.003 Culture.Performing arts 0.021 0.003 Culture.Philosophy and religion 0.042 0.01 Culture.Sports 0.093 0.06 Culture.Visual arts.Architecture 0.041 0.011 Culture.Visual arts.Comics and Anime 0.024 0.002 Culture.Visual arts.Fashion 0.019 0.001 Culture.Visual arts.Visual arts* 0.095 0.018 Geography.Geographical 0.054 0.021 Geography.Regions.Africa.Africa* 0.101 0.008 Geography.Regions.Africa.Central Africa 0.018 0.001 Geography.Regions.Africa.Eastern Africa 0.017 0.001 Geography.Regions.Africa.Northern Africa 0.02 0.001 Geography.Regions.Africa.Southern Africa 0.019 0.001 Geography.Regions.Africa.Western Africa 0.018 0.001 Geography.Regions.Americas.Central America 0.021 0.003 Geography.Regions.Americas.North America 0.119 0.063 Geography.Regions.Americas.South America 0.024 0.007 Geography.Regions.Asia.Asia* 0.182 0.052 Geography.Regions.Asia.Central Asia 0.017 0.001 Geography.Regions.Asia.East Asia 0.042 0.012 Geography.Regions.Asia.North Asia 0.032 0.006 Geography.Regions.Asia.South Asia 0.037 0.016 Geography.Regions.Asia.Southeast Asia 0.027 0.006 Geography.Regions.Asia.West Asia 0.034 0.012 Geography.Regions.Europe.Eastern Europe 0.055 0.018 Geography.Regions.Europe.Europe* 0.202 0.081 Geography.Regions.Europe.Northern Europe 0.066 0.029 Geography.Regions.Europe.Southern Europe 0.038 0.014 Geography.Regions.Europe.Western Europe 0.048 0.02 Geography.Regions.Oceania 0.041 0.016 History and Society.Business and economics 0.055 0.01 History and Society.Education 0.035 0.008 History and Society.History 0.05 0.011 History and Society.Military and warfare 0.051 0.015 History and Society.Politics and government 0.072 0.028 History and Society.Society 0.046 0.008 History and Society.Transportation 0.057 0.016 STEM.Biology 0.046 0.034 STEM.Chemistry 0.02 0.002 STEM.Computing 0.031 0.003 STEM.Earth and environment 0.025 0.005 STEM.Engineering 0.034 0.006 STEM.Libraries & Information 0.018 0.001 STEM.Mathematics 0.018 0 STEM.Medicine & Health 0.027 0.006 STEM.Physics 0.019 0.001 STEM.STEM* 0.257 0.065 STEM.Space 0.021 0.004 STEM.Technology 0.057 0.005

Test data performance:

Test data performance
Label Match rate Filter rate Recall Precision f1 Accuracy ROC AUC PR AUC Culture.Biography.Biography* 0.122 0.878 0.946 0.929 0.937 0.985 0.982 0.952 Culture.Biography.Women 0.023 0.977 0.76 0.504 0.606 0.985 0.975 0.589 Culture.Food and drink 0.003 0.997 0.465 0.371 0.413 0.997 0.937 0.352 Culture.Internet culture 0.005 0.995 0.657 0.517 0.578 0.996 0.96 0.549 Culture.Linguistics 0.006 0.994 0.637 0.852 0.729 0.996 0.954 0.656 Culture.Literature 0.018 0.982 0.745 0.619 0.676 0.989 0.965 0.726 Culture.Media.Books 0.006 0.994 0.79 0.609 0.688 0.997 0.971 0.659 Culture.Media.Entertainment 0.005 0.995 0.495 0.429 0.459 0.995 0.947 0.433 Culture.Media.Films 0.011 0.989 0.826 0.829 0.828 0.996 0.974 0.813 Culture.Media.Media* 0.066 0.934 0.805 0.67 0.731 0.968 0.966 0.813 Culture.Media.Music 0.02 0.98 0.785 0.807 0.795 0.992 0.974 0.818 Culture.Media.Radio 0.002 0.998 0.741 0.711 0.726 0.999 0.962 0.741 Culture.Media.Software 0.005 0.995 0.391 0.094 0.152 0.994 0.946 0.094 Culture.Media.Television 0.009 0.991 0.677 0.68 0.678 0.994 0.964 0.664 Culture.Media.Video games 0.003 0.997 0.819 0.732 0.773 0.999 0.977 0.801 Culture.Performing arts 0.004 0.996 0.555 0.478 0.514 0.997 0.947 0.414 Culture.Philosophy and religion 0.009 0.991 0.397 0.472 0.431 0.989 0.91 0.339 Culture.Sports 0.057 0.943 0.875 0.929 0.901 0.988 0.976 0.933 Culture.Visual arts.Architecture 0.011 0.989 0.705 0.672 0.688 0.993 0.969 0.673 Culture.Visual arts.Comics and Anime 0.004 0.996 0.668 0.416 0.513 0.997 0.966 0.558 Culture.Visual arts.Fashion 0.002 0.998 0.558 0.242 0.338 0.998 0.95 0.215 Culture.Visual arts.Visual arts* 0.022 0.978 0.681 0.566 0.618 0.985 0.952 0.666 Geography.Geographical 0.019 0.981 0.643 0.7 0.67 0.987 0.956 0.698 Geography.Regions.Africa.Africa* 0.013 0.987 0.723 0.462 0.564 0.991 0.96 0.639 Geography.Regions.Africa.Central Africa 0.002 0.998 0.609 0.244 0.349 0.998 0.954 0.321 Geography.Regions.Africa.Eastern Africa 0.001 0.999 0.632 0.262 0.371 0.999 0.957 0.253 Geography.Regions.Africa.Northern Africa 0.003 0.997 0.605 0.321 0.42 0.998 0.945 0.343 Geography.Regions.Africa.Southern Africa 0.002 0.998 0.691 0.411 0.515 0.998 0.959 0.514 Geography.Regions.Africa.Western Africa 0.002 0.998 0.678 0.297 0.413 0.999 0.96 0.277 Geography.Regions.Americas.Central America 0.003 0.997 0.531 0.569 0.55 0.997 0.932 0.494 Geography.Regions.Americas.North America 0.061 0.939 0.664 0.682 0.673 0.959 0.949 0.726 Geography.Regions.Americas.South America 0.007 0.993 0.706 0.681 0.693 0.996 0.963 0.691 Geography.Regions.Asia.Asia* 0.053 0.947 0.724 0.715 0.719 0.97 0.949 0.756 Geography.Regions.Asia.Central Asia 0.002 0.998 0.618 0.306 0.41 0.999 0.952 0.462 Geography.Regions.Asia.East Asia 0.012 0.988 0.636 0.665 0.65 0.992 0.95 0.625 Geography.Regions.Asia.North Asia 0.006 0.994 0.644 0.579 0.609 0.995 0.946 0.55 Geography.Regions.Asia.South Asia 0.013 0.987 0.681 0.839 0.752 0.993 0.953 0.708 Geography.Regions.Asia.Southeast Asia 0.006 0.994 0.615 0.668 0.641 0.996 0.942 0.557 Geography.Regions.Asia.West Asia 0.01 0.99 0.682 0.794 0.734 0.994 0.953 0.662 Geography.Regions.Europe.Eastern Europe 0.017 0.983 0.7 0.771 0.733 0.991 0.951 0.71 Geography.Regions.Europe.Europe* 0.091 0.909 0.724 0.642 0.681 0.945 0.943 0.744 Geography.Regions.Europe.Northern Europe 0.027 0.973 0.609 0.643 0.626 0.979 0.948 0.644 Geography.Regions.Europe.Southern Europe 0.013 0.987 0.642 0.673 0.657 0.991 0.949 0.618 Geography.Regions.Europe.Western Europe 0.02 0.98 0.629 0.657 0.643 0.986 0.95 0.63 Geography.Regions.Oceania 0.014 0.986 0.705 0.839 0.766 0.993 0.96 0.75 History and Society.Business and economics 0.014 0.986 0.441 0.315 0.367 0.985 0.936 0.248 History and Society.Education 0.008 0.992 0.496 0.489 0.493 0.992 0.944 0.4 History and Society.History 0.01 0.99 0.364 0.403 0.382 0.987 0.916 0.315 History and Society.Military and warfare 0.013 0.987 0.518 0.622 0.565 0.988 0.933 0.521 History and Society.Politics and government 0.02 0.98 0.524 0.731 0.611 0.981 0.925 0.603 History and Society.Society 0.005 0.995 0.302 0.48 0.371 0.992 0.871 0.318 History and Society.Transportation 0.014 0.986 0.721 0.809 0.762 0.993 0.963 0.712 STEM.Biology 0.028 0.972 0.767 0.948 0.848 0.991 0.962 0.816 STEM.Chemistry 0.003 0.997 0.543 0.294 0.382 0.997 0.958 0.27 STEM.Computing 0.007 0.993 0.421 0.182 0.254 0.993 0.951 0.149 STEM.Earth and environment 0.004 0.996 0.564 0.594 0.579 0.996 0.947 0.522 STEM.Engineering 0.006 0.994 0.585 0.596 0.591 0.995 0.947 0.51 STEM.Libraries & Information 0.002 0.998 0.515 0.203 0.291 0.998 0.948 0.238 STEM.Mathematics 0.002 0.998 0.27 0.068 0.109 0.998 0.942 0.125 STEM.Medicine & Health 0.006 0.994 0.446 0.499 0.471 0.994 0.933 0.398 STEM.Physics 0.002 0.998 0.368 0.168 0.23 0.998 0.945 0.126 STEM.STEM* 0.104 0.896 0.767 0.477 0.588 0.93 0.955 0.768 STEM.Space 0.004 0.996 0.683 0.795 0.735 0.998 0.963 0.686 STEM.Technology 0.009 0.991 0.383 0.219 0.279 0.99 0.928 0.213

Implementation

Model architecture

Model architecture
{ "type": "GradientBoosting", "params": { "scale": false, "center": false, "labels": [ "Culture.Biography.Biography*", "Culture.Biography.Women", "Culture.Food and drink", "Culture.Internet culture", "Culture.Linguistics", "Culture.Literature", "Culture.Media.Books", "Culture.Media.Entertainment", "Culture.Media.Films", "Culture.Media.Media*", "Culture.Media.Music", "Culture.Media.Radio", "Culture.Media.Software", "Culture.Media.Television", "Culture.Media.Video games", "Culture.Performing arts", "Culture.Philosophy and religion", "Culture.Sports", "Culture.Visual arts.Architecture", "Culture.Visual arts.Comics and Anime", "Culture.Visual arts.Fashion", "Culture.Visual arts.Visual arts*", "Geography.Geographical", "Geography.Regions.Africa.Africa*", "Geography.Regions.Africa.Central Africa", "Geography.Regions.Africa.Eastern Africa", "Geography.Regions.Africa.Northern Africa", "Geography.Regions.Africa.Southern Africa", "Geography.Regions.Africa.Western Africa", "Geography.Regions.Americas.Central America", "Geography.Regions.Americas.North America", "Geography.Regions.Americas.South America", "Geography.Regions.Asia.Asia*", "Geography.Regions.Asia.Central Asia", "Geography.Regions.Asia.East Asia", "Geography.Regions.Asia.North Asia", "Geography.Regions.Asia.South Asia", "Geography.Regions.Asia.Southeast Asia", "Geography.Regions.Asia.West Asia", "Geography.Regions.Europe.Eastern Europe", "Geography.Regions.Europe.Europe*", "Geography.Regions.Europe.Northern Europe", "Geography.Regions.Europe.Southern Europe", "Geography.Regions.Europe.Western Europe", "Geography.Regions.Oceania", "History and Society.Business and economics", "History and Society.Education", "History and Society.History", "History and Society.Military and warfare", "History and Society.Politics and government", "History and Society.Society", "History and Society.Transportation", "STEM.Biology", "STEM.Chemistry", "STEM.Computing", "STEM.Earth and environment", "STEM.Engineering", "STEM.Libraries & Information", "STEM.Mathematics", "STEM.Medicine & Health", "STEM.Physics", "STEM.STEM*", "STEM.Space", "STEM.Technology" ], "multilabel": true, "population_rates": null, "ccp_alpha": 0.0, "criterion": "friedman_mse", "init": null, "learning_rate": 0.1, "loss": "deviance", "max_depth": 5, "max_features": "log2", "max_leaf_nodes": null, "min_impurity_decrease": 0.0, "min_impurity_split": null, "min_samples_leaf": 1, "min_samples_split": 2, "min_weight_fraction_leaf": 0.0, "n_estimators": 150, "n_iter_no_change": null, "presort": "deprecated", "random_state": null, "subsample": 1.0, "tol": 0.0001, "validation_fraction": 0.1, "verbose": 0, "warm_start": false, "label_weights": {} } }

Output schema

Output schema
{ "title": "Scikit learn-based classifier score with probability", "type": "object", "properties": { "prediction": { "description": "The most likely labels predicted by the estimator", "type": "array", "items": { "type": "string" } }, "probability": { "description": "A mapping of probabilities onto each of the potential output labels", "type": "object", "properties": { "Culture.Biography.Biography*": { "type": "number" }, "Culture.Biography.Women": { "type": "number" }, "Culture.Food and drink": { "type": "number" }, "Culture.Internet culture": { "type": "number" }, "Culture.Linguistics": { "type": "number" }, "Culture.Literature": { "type": "number" }, "Culture.Media.Books": { "type": "number" }, "Culture.Media.Entertainment": { "type": "number" }, "Culture.Media.Films": { "type": "number" }, "Culture.Media.Media*": { "type": "number" }, "Culture.Media.Music": { "type": "number" }, "Culture.Media.Radio": { "type": "number" }, "Culture.Media.Software": { "type": "number" }, "Culture.Media.Television": { "type": "number" }, "Culture.Media.Video games": { "type": "number" }, "Culture.Performing arts": { "type": "number" }, "Culture.Philosophy and religion": { "type": "number" }, "Culture.Sports": { "type": "number" }, "Culture.Visual arts.Architecture": { "type": "number" }, "Culture.Visual arts.Comics and Anime": { "type": "number" }, "Culture.Visual arts.Fashion": { "type": "number" }, "Culture.Visual arts.Visual arts*": { "type": "number" }, "Geography.Geographical": { "type": "number" }, "Geography.Regions.Africa.Africa*": { "type": "number" }, "Geography.Regions.Africa.Central Africa": { "type": "number" }, "Geography.Regions.Africa.Eastern Africa": { "type": "number" }, "Geography.Regions.Africa.Northern Africa": { "type": "number" }, "Geography.Regions.Africa.Southern Africa": { "type": "number" }, "Geography.Regions.Africa.Western Africa": { "type": "number" }, "Geography.Regions.Americas.Central America": { "type": "number" }, "Geography.Regions.Americas.North America": { "type": "number" }, "Geography.Regions.Americas.South America": { "type": "number" }, "Geography.Regions.Asia.Asia*": { "type": "number" }, "Geography.Regions.Asia.Central Asia": { "type": "number" }, "Geography.Regions.Asia.East Asia": { "type": "number" }, "Geography.Regions.Asia.North Asia": { "type": "number" }, "Geography.Regions.Asia.South Asia": { "type": "number" }, "Geography.Regions.Asia.Southeast Asia": { "type": "number" }, "Geography.Regions.Asia.West Asia": { "type": "number" }, "Geography.Regions.Europe.Eastern Europe": { "type": "number" }, "Geography.Regions.Europe.Europe*": { "type": "number" }, "Geography.Regions.Europe.Northern Europe": { "type": "number" }, "Geography.Regions.Europe.Southern Europe": { "type": "number" }, "Geography.Regions.Europe.Western Europe": { "type": "number" }, "Geography.Regions.Oceania": { "type": "number" }, "History and Society.Business and economics": { "type": "number" }, "History and Society.Education": { "type": "number" }, "History and Society.History": { "type": "number" }, "History and Society.Military and warfare": { "type": "number" }, "History and Society.Politics and government": { "type": "number" }, "History and Society.Society": { "type": "number" }, "History and Society.Transportation": { "type": "number" }, "STEM.Biology": { "type": "number" }, "STEM.Chemistry": { "type": "number" }, "STEM.Computing": { "type": "number" }, "STEM.Earth and environment": { "type": "number" }, "STEM.Engineering": { "type": "number" }, "STEM.Libraries & Information": { "type": "number" }, "STEM.Mathematics": { "type": "number" }, "STEM.Medicine & Health": { "type": "number" }, "STEM.Physics": { "type": "number" }, "STEM.STEM*": { "type": "number" }, "STEM.Space": { "type": "number" }, "STEM.Technology": { "type": "number" } } } } }

Example input and output

Input:

https://ores.wikimedia.org/v3/scores/wikidatawiki/1907686315/itemtopic

Output:

Example output
{ "wikidatawiki": { "models": { "itemtopic": { "version": "1.2.0" } }, "scores": { "1907686315": { "itemtopic": { "score": { "prediction": [ "STEM.STEM*" ], "probability": { "Culture.Biography.Biography*": 0.009059893345632097, "Culture.Biography.Women": 0.0006924491258526178, "Culture.Food and drink": 0.0006399242658997215, "Culture.Internet culture": 0.0009384780459913412, "Culture.Linguistics": 0.0018606277391225432, "Culture.Literature": 0.003990388751181737, "Culture.Media.Books": 0.0006214752106656115, "Culture.Media.Entertainment": 0.001104834881085509, "Culture.Media.Films": 0.0011465477594696284, "Culture.Media.Media*": 0.009497882960118977, "Culture.Media.Music": 0.0005314326820035878, "Culture.Media.Radio": 0.0001418663128807519, "Culture.Media.Software": 0.0006122966374525156, "Culture.Media.Television": 0.0011153877562536376, "Culture.Media.Video games": 0.0006372239889671269, "Culture.Performing arts": 0.0006531356388159476, "Culture.Philosophy and religion": 0.01399521934257544, "Culture.Sports": 0.0018462250677368348, "Culture.Visual arts.Architecture": 0.0016560396166840437, "Culture.Visual arts.Comics and Anime": 0.0005305955236163667, "Culture.Visual arts.Fashion": 0.000537788411976724, "Culture.Visual arts.Visual arts*": 0.009907875401930734, "Geography.Geographical": 0.01571363482516823, "Geography.Regions.Africa.Africa*": 0.020280349224975614, "Geography.Regions.Africa.Central Africa": 0.0007006250310735848, "Geography.Regions.Africa.Eastern Africa": 0.000981468869640802, "Geography.Regions.Africa.Northern Africa": 0.015712323087656205, "Geography.Regions.Africa.Southern Africa": 0.001221937118377821, "Geography.Regions.Africa.Western Africa": 0.0008305320623083369, "Geography.Regions.Americas.Central America": 0.001306842712476455, "Geography.Regions.Americas.North America": 0.030570993625411366, "Geography.Regions.Americas.South America": 0.009381192562807516, "Geography.Regions.Asia.Asia*": 0.08763779333893186, "Geography.Regions.Asia.Central Asia": 0.0021630529281042718, "Geography.Regions.Asia.East Asia": 0.01090968773383821, "Geography.Regions.Asia.North Asia": 0.029951228290233667, "Geography.Regions.Asia.South Asia": 0.005977584426712786, "Geography.Regions.Asia.Southeast Asia": 0.0028688045552628266, "Geography.Regions.Asia.West Asia": 0.0009526856502617891, "Geography.Regions.Europe.Eastern Europe": 0.029972291587851183, "Geography.Regions.Europe.Europe*": 0.13296776378542635, "Geography.Regions.Europe.Northern Europe": 0.016907973275604154, "Geography.Regions.Europe.Southern Europe": 0.005813048163270592, "Geography.Regions.Europe.Western Europe": 0.005037055635498127, "Geography.Regions.Oceania": 0.007780720915153282, "History and Society.Business and economics": 0.005890874106250135, "History and Society.Education": 0.0017680572320172617, "History and Society.History": 0.01973006391755843, "History and Society.Military and warfare": 0.006573635883462243, "History and Society.Politics and government": 0.007573132449112524, "History and Society.Society": 0.04381007914549254, "History and Society.Transportation": 0.002797769913886188, "STEM.Biology": 0.005780672890531569, "STEM.Chemistry": 0.0022570835539507676, "STEM.Computing": 0.0018290751421398967, "STEM.Earth and environment": 0.0795914195853073, "STEM.Engineering": 0.004058097854564882, "STEM.Libraries & Information": 0.0010339015208737487, "STEM.Mathematics": 0.0017040157655581244, "STEM.Medicine & Health": 0.005650365932513206, "STEM.Physics": 0.020150498627265184, "STEM.STEM*": 0.8790296717258461, "STEM.Space": 0.11458869168317454, "STEM.Technology": 0.012381701761546463 } } } } } } }

Data

Data pipeline

The data to train was fetched from a set of revision IDs. Then various pieces of information about the revision were extracted using automated processes, and the revision text was fed into word2vec to get an item embedding. Finally, labels are derived from the mid-level WikiProject categories that the item is associated with.

Training data

Training data was automatically and randomly separated from test data during training using the drafttopic git repository (which trains both drafttopic and articletopic models).

Test data

Test data was automatically and randomly split off from train data using the drafttopic git repository (which trains both drafttopic and articletopic models). The model then makes a prediction on that data, which is compared to the underlying ground truth to calculate performance statistics.

Licenses

• Code: MIT license
• Model: MIT license

Citation

Cite this model card as:

@misc{
  Triedman_Bazira_2023_Wikidata_item_topic,
  title={ Wikidata item topic model card },
  author={ Triedman, Harold and Bazira, Kevin },
  year={ 2023 },
  url={ https://meta.wikimedia.org/wiki/Machine_learning_models/Production/Wikidata_item_topic }
}