Hugging Face's logo

Datasets:
Feng-001
/
Clustering-Benchmark

Tasks:
Tabular Classification
Languages:
English
Size:
100K<n<1M
License:
Dataset card Files
xet
Community
Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

YAML Metadata Warning: The task_categories "tabular-clustering" is not in the official list: text-classification, token-classification, table-question-answering, question-answering, zero-shot-classification, translation, summarization, feature-extraction, text-generation, fill-mask, sentence-similarity, text-to-speech, text-to-audio, automatic-speech-recognition, audio-to-audio, audio-classification, audio-text-to-text, voice-activity-detection, depth-estimation, image-classification, object-detection, image-segmentation, text-to-image, image-to-text, image-to-image, image-to-video, unconditional-image-generation, video-classification, reinforcement-learning, robotics, tabular-classification, tabular-regression, tabular-to-text, table-to-text, multiple-choice, text-ranking, text-retrieval, time-series-forecasting, text-to-video, image-text-to-text, image-text-to-image, image-text-to-video, visual-question-answering, document-question-answering, zero-shot-image-classification, graph-ml, mask-generation, zero-shot-object-detection, text-to-3d, image-to-3d, image-feature-extraction, video-text-to-text, keypoint-detection, visual-document-retrieval, any-to-any, video-to-video, other

CLUBench: A Clustering Benchmark

This repository provides CLUBench, a comprehensive benchmark for tabular clustering.

Paper: [CLUBench: A Clustering Benchmark]

Code: CLUBench

Abstract

Clustering is a fundamental problem in data science with a long-standing research history. Over the past decades, numerous clustering algorithms, ranging from conventional machine learning approaches to deep clustering methods, have been developed. Despite this progress, a systematic and large-scale empirical evaluation that jointly considers conventional algorithms, deep learning-based methods and recent foundation model-based clustering remains largely absent, leading to limited guidance on algorithm selection and deployment. To address this gap, we introduce CLUBench, a comprehensive clustering benchmark comprising 24 algorithms of diverse principles evaluated on 131 datasets across tabular, text and image data. Importantly, CLUBench provides a unified comparison between state-of-the-art baselines and foundation model-energized clustering strategies on all three modalities (tabular, text and image). Extensive experiments (178,815) in CLUBench yield statistically meaningful insights and identify promising yet underexplored pathways about clustering research. For example, we observe low-rank structure in cross-model performance matrices, which facilitates an efficient strategy for rapid algorithm evaluation and selection in practical applications. In addition, we provide an easy-to-use toolbox by encapsulating the source codes from the official code repository into a unified framework, accompanied by detailed instructions.

Dataset Details

Source From:

ID Dataset Type size dim clusters r_mm r_ma IR
[1] echocardiogram tabular 61 10 2 0.386 0.279 0.221
[2] skillcraft1_master_table_dataset tabular 3303 18 6 0.206 0.051 0.071
[3] breast_cancer_wisconsin_original tabular 683 9 2 0.538 0.350 0.150
[4] smoker_condition tabular 1012 7 2 0.656 0.396 0.104
[5] glass_identification tabular 214 9 6 0.118 0.042 0.127
[6] statlog_image_segmentation tabular 2310 19 7 1.000 0.143 0.000
[7] planning_relax tabular 182 12 2 0.400 0.286 0.214
[8] customer_classification tabular 1000 11 4 0.772 0.217 0.025
[9] pima_indians_diabetes_database tabular 768 8 2 0.536 0.349 0.151
[10] mobile_price_classification tabular 2000 20 4 1.000 0.250 0.000
[11] spambase tabular 4601 57 2 0.650 0.394 0.106
[12] rice_seed_gonen_jasmine tabular 9999 10 2 0.821 0.451 0.049
[13] heart_attack_analysis_prediction_dataset tabular 303 13 2 0.836 0.455 0.045
[14] user_knowledge_modeling tabular 258 5 4 0.273 0.093 0.098
[15] world12d tabular 150 12 5 0.190 0.053 0.088
[16] pumpkin_seeds tabular 2500 12 2 0.923 0.480 0.020
[17] iris tabular 150 4 3 1.000 0.333 0.000
[18] wine tabular 178 13 3 0.676 0.270 0.053
[19] letter_recognition tabular 9992 16 26 0.904 0.037 0.001
[20] mammographic_mass tabular 830 5 2 0.944 0.486 0.014
[21] breast_tissue tabular 106 9 6 0.636 0.132 0.028
[22] hepatitis tabular 80 19 2 0.194 0.163 0.338
[23] predicting_pulsar_star tabular 9273 8 2 0.101 0.092 0.408
[24] breast_cancer_wisconsin_prognostic tabular 569 30 2 0.594 0.373 0.127
[25] wireless_indoor_localization tabular 2000 7 4 1.000 0.250 0.000
[26] date_fruit tabular 898 34 7 0.319 0.072 0.062
[27] zoo tabular 101 16 7 0.098 0.040 0.118
[28] htru2 tabular 9999 8 2 0.101 0.092 0.408
[29] ionosphere tabular 351 34 2 0.560 0.359 0.141
[30] music_genre_classification tabular 1000 26 10 1.000 0.100 0.000
[31] spectf_heart tabular 80 44 2 1.000 0.500 0.000
[32] rice_dataset_cammeo_and_osmancik tabular 3810 7 2 0.748 0.428 0.072
[33] ph_recognition tabular 653 3 15 0.864 0.058 0.002
[34] banknote_authentication tabular 1372 4 2 0.801 0.445 0.055
[35] wine_quality tabular 4873 11 5 0.074 0.033 0.160
[36] cardiovascular_study tabular 2927 15 2 0.179 0.152 0.348
[37] statlog_german_credit tabular 1000 24 2 0.429 0.300 0.200
[38] boston tabular 154 13 3 0.371 0.169 0.121
[39] seismic_bumps tabular 646 24 2 0.071 0.067 0.433
[40] dry_bean tabular 9997 16 7 0.147 0.038 0.065
[41] credit_risk_classification tabular 976 11 2 0.239 0.193 0.307
[42] epileptic_seizure_recognition tabular 5750 178 5 1.000 0.200 0.000
[43] website_phishing tabular 1353 9 3 0.147 0.076 0.188
[44] optical_recognition_of_handwritten_digits tabular 3823 64 10 0.967 0.098 0.001
[45] siberian_weather_stats tabular 1407 11 7 0.073 0.024 0.122
[46] orbit_classification_for_prediction_nasa tabular 1722 11 3 0.065 0.056 0.371
[47] magic_gamma_telescope tabular 9999 10 2 0.542 0.352 0.148
[48] raisin tabular 900 7 2 1.000 0.500 0.000
[49] patient_treatment_classification tabular 4412 10 2 0.679 0.404 0.096
[50] fetal_health_classification tabular 2126 21 3 0.106 0.083 0.316
[51] dermatology tabular 358 34 6 0.180 0.056 0.373
[52] secom tabular 1567 590 2 0.071 0.066 0.000
[53] paris_housing_classification tabular 10000 17 2 0.145 0.127 0.053
[54] seeds tabular 210 7 3 1.000 0.333 0.275
[55] wine_customer tabular 178 13 3 0.676 0.270 0.000
[56] crowdsourced_mapping tabular 9997 28 4 0.060 0.043 0.212
[57] durum_wheat_features tabular 9000 236 3 1.000 0.333 0.093
[58] classification_in_asteroseismology tabular 1001 3 2 0.404 0.288 0.063
[59] birds_bones_and_living_habits tabular 413 10 6 0.185 0.056 0.000
[60] microbes tabular 9995 24 10 0.082 0.020 0.097
[61] image_segmentation tabular 210 19 7 1.000 0.143 0.440
[62] water_quality tabular 2011 9 2 0.676 0.403 0.235
[63] insurance_company_benchmark tabular 5822 85 2 0.064 0.060 0.115
[64] harbermans_survival tabular 306 3 2 0.360 0.265 0.175
[65] yeast tabular 1459 8 8 0.065 0.021 0.132
[66] heart_disease tabular 297 13 5 0.081 0.044 0.004
[67] ecoli tabular 327 7 5 0.140 0.061 0.052
[68] extyaleb tabular 319 30 5 0.954 0.194 0.171
[69] breast_cancer_coimbra tabular 116 9 2 0.812 0.448 0.061
[70] student_grade tabular 395 29 2 0.491 0.329 0.234
[71] human_stress_detection tabular 2001 3 3 0.634 0.250 0.004
[72] fraud_detection_bank tabular 9999 112 2 0.362 0.266 0.031
[73] pen_based_recognition_of_handwritten_digits tabular 7494 16 10 0.922 0.096 0.000
[74] diabetic_retinopathy_debrecen tabular 1151 19 2 0.884 0.469 0.026
[75] pistachio tabular 2148 28 2 0.744 0.426 0.262
[76] turkish_music_emotion tabular 400 50 4 1.000 0.250 0.000
[77] parkinsons tabular 195 22 2 0.327 0.246 0.000
[78] weather tabular 365 192 7 0.603 0.121 0.148
[79] blood_transfusion_service_center tabular 748 4 2 0.312 0.238 0.004
[80] mfeat-karhunen tabular 2000 64 10 1.000 0.100 0.039
[81] mfeat-factors tabular 2000 216 10 1.000 0.100 0.116
[82] wall-robot-navigation tabular 5456 24 4 0.149 0.060 0.007
[83] Waveform tabular 5000 21 3 0.971 0.329 0.053
[84] gas-drift tabular 10000 128 6 0.546 0.118 0.005
[85] mfeat-morphological tabular 2000 6 10 1.000 0.100 0.000
[86] JapaneseVowels tabular 9961 14 9 0.485 0.079 0.124
[87] rmftsa_sleepdata tabular 1024 2 4 0.233 0.092 0.337
[88] first-order-theorem-proving tabular 6118 51 6 0.190 0.079 0.062
[89] gina_prior2 tabular 3468 784 10 0.822 0.091 0.153
[90] fabert tabular 8237 800 7 0.261 0.061 0.064
[91] dilbert tabular 10000 2000 5 0.934 0.191 0.000
[92] synthetic_control tabular 600 60 6 1.000 0.167 0.009
[93] Drug Consumption tabular 1749 12 4 0.261 0.113 0.053
[94] shuttle tabular 10000 9 2 0.195 0.163 0.005
[95] tr45.wc tabular 676 8261 9 0.113 0.027 0.000
[96] steel-plates-fault tabular 1941 33 2 0.531 0.347 0.000
[97] fbis.wc tabular 2196 2000 11 0.128 0.030 0.000
[98] mfeat-fourier tabular 2000 76 10 1.000 0.100 0.000
[99] vehicle tabular 846 18 4 0.913 0.235 0.000
[100] micro-mass tabular 360 1300 10 1.000 0.100 0.000
[101] ISOLET tabular 7797 617 26 0.993 0.038 0.000
[102] poker-hand tabular 10000 10 2 0.843 0.457 0.000
[103] tamilnadu-electricity tabular 10000 2 20 0.480 0.030 0.000
[104] mnist64 image 1082 64 6 0.967 0.164 0.000
[105] MNIST_CLIP^+ image 9996 512 10 0.801 0.090 0.000
[106] fashion_mnist image 3000 784 10 1.000 0.100 0.001
[107] FashionMNIST_CLIP^+ image 10000 512 10 1.000 0.100 0.038
[108] cifar10 image 3250 1024 10 1.000 0.100 0.021
[109] CIFAR10_CLIP^+ image 10000 512 10 1.000 0.100 0.153
[110] coil20^* image 1440 400 20 1.000 0.050 0.062
[111] COIL20_CLIP^+ image 1440 512 20 1.000 0.050 0.000
[112] labeled_faces_in_the_wild image 2200 5828 2 1.000 0.500 0.006
[113] flickr_material_database image 997 1536 10 0.990 0.099 0.053
[114] street_view_house_numbers image 732 1024 10 0.341 0.064 0.000
[115] har image 735 561 6 0.702 0.135 0.006
[116] Indian_pines image 8858 220 5 0.121 0.055 0.000
[117] satellite_image image 6435 36 6 0.408 0.097 0.000
[118] olivetti_faces image 400 4096 40 1.000 0.025 0.000
[119] cnae9 text 1080 856 9 1.000 0.111 0.000
[120] imdb text 3250 700 2 1.000 0.500 0.000
[121] hate_speech text 3221 100 3 0.075 0.058 0.000
[122] sentiment_labeld_sentences text 2748 200 2 0.983 0.496 0.000
[123] sms_spam_collection text 835 500 2 0.155 0.134 0.000
[124] wos text 9997 4096 7 0.223 0.069 0.000
[125] enron text 9999 4096 2 0.990 0.497 0.315
[126] reuters text 6576 4096 3 0.562 0.243 0.004
[127] 20newsgroups text 9991 4096 20 0.612 0.033 0.366
[128] Mouse_retina tabular (BioInfo) 8352 6198 5 0.054 0.043 0.073
[129] Campbell tabular (BioInfo) 9993 26774 14 0.052 0.024 0.003
[130] PCam image 4000 27648 2 0.977 0.494 0.302
[131] Baron Human tabular (BioInfo) 8451 20125 9 0.069 0.020 0.111
  • Curated by: Feng Xiao, Dazhi Fu, Jicong Fan
  • Language(s) (NLP): English
  • License: Apache-2,0
Downloads last month
9