{"id":"W4415097369","doi":"10.48550/arxiv.2510.06840","title":"CNN-TFT explained by SHAP with multi-head attention weights for time series forecasting","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Neural Networks and Applications","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Fundação para a Ciência e a Tecnologia; Natural Sciences and Engineering Research Council of Canada","keywords":"Time series; Convolutional neural network; Salient; Pattern recognition (psychology); Multivariate statistics; Transformer; Feature extraction; Deep learning; Series (stratigraphy)","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004846739,0.000822518,0.0005158591,0.0005665962,0.0002596789,0.0005980594,0.0009328369,0.0007997673,0.002460443],"category_scores_gemma":[0.00183214,0.0003339063,0.0008827409,0.0007306607,0.000356859,0.001561494,0.0008147536,0.0012014,0.0004253969],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008802002,"about_ca_system_score_gemma":0.0008606309,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0115483,"about_ca_topic_score_gemma":0.01140578,"domain_scores_codex":[0.9998524,0.00002291453,0.00001025951,0.00005589386,0.00003298308,0.00002550774],"domain_scores_gemma":[0.9997008,0.0001005861,0.00004268697,0.00006220141,0.00006896753,0.00002479861],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000147685,0.00005879109,0.00265564,0.0000600486,0.0001187993,0.0001636405,0.00006799896,0.7727193,0.008228059,0.01710614,0.004761902,0.1939119],"study_design_scores_gemma":[0.000002028749,0.000007523224,0.0001339497,0.000001375916,0.000007085068,0.000009033363,0.00000192696,0.9953266,0.0005663503,0.003740781,0.0002009486,0.000002263423],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.07387017,0.0006146851,0.9192854,0.0005025902,0.0001325762,0.00003199056,0.000430969,0.002501071,0.002630484],"genre_scores_gemma":[0.9148816,0.0003374382,0.07928906,0.000187178,0.00009601244,0.00004621137,0.0008657245,0.0001318315,0.004165013],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0115483,"threshold_uncertainty_score":0.02296221,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05618668288287484,"score_gpt":0.2755162724223891,"score_spread":0.2193295895395142,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}