{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,3]],"date-time":"2026-04-03T14:55:48Z","timestamp":1775228148860,"version":"3.50.1"},"reference-count":35,"publisher":"Springer Science and Business Media LLC","issue":"2","license":[{"start":{"date-parts":[[2020,5,20]],"date-time":"2020-05-20T00:00:00Z","timestamp":1589932800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2020,5,20]],"date-time":"2020-05-20T00:00:00Z","timestamp":1589932800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100001691","name":"Japan Society for the Promotion of Science","doi-asserted-by":"publisher","award":["JP17J04768"],"award-info":[{"award-number":["JP17J04768"]}],"id":[{"id":"10.13039\/501100001691","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Data Sci. Eng."],"published-print":{"date-parts":[[2020,6]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:p>Although deep neural networks are excellent for text sentiment analysis, their applications in real-world practice are occasionally limited owing to their black-box property. In this study, we propose a novel neural network model called contextual sentiment neural network (CSNN) model that can explain the process of its sentiment analysis prediction in a way that humans find natural and agreeable and can catch up the summary of the contents. The CSNN has the following interpretable layers: the word-level original sentiment layer, word-level sentiment shift layer, word-level global importance layer, word-level contextual sentiment layer, and concept-level contextual sentiment layer. Because of these layers, this network can explain the process of its document-level sentiment analysis results in a human-like way using these layers. Realizing the interpretability of each layer in the CSNN is a crucial problem in the development of this CSNN because the general back-propagation method cannot realize such interpretability. To realize this interpretability, we propose a novel learning strategy called initialization propagation (IP) learning. Using real textual datasets, we experimentally demonstrate that the proposed IP learning is effective for improving the interpretability of each layer in CSNN. We then experimentally demonstrate that the CSNN has both the high predictability and high explanation ability.<\/jats:p>","DOI":"10.1007\/s41019-020-00122-4","type":"journal-article","created":{"date-parts":[[2020,5,20]],"date-time":"2020-05-20T12:02:55Z","timestamp":1589976175000},"page":"180-192","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":36,"title":["Contextual Sentiment Neural Network for Document Sentiment Analysis"],"prefix":"10.1007","volume":"5","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-4200-1311","authenticated-orcid":false,"given":"Tomoki","family":"Ito","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kota","family":"Tsubouchi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hiroki","family":"Sakaji","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Tatsuo","family":"Yamashita","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kiyoshi","family":"Izumi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"297","published-online":{"date-parts":[[2020,5,20]]},"reference":[{"key":"122_CR1","doi-asserted-by":"crossref","unstructured":"Kim Y (2014) Convolutional neural networks for sentence classification. In: EMNLP 2014","DOI":"10.3115\/v1\/D14-1181"},{"key":"122_CR2","first-page":"129","volume":"2013","author":"S Li","year":"2013","unstructured":"Li S, Wang Z, Lee SYM, Huang C-R (2013) Sentiment classification with polarity shifting detection. IALP 2013:129\u2013132","journal-title":"IALP"},{"key":"122_CR3","first-page":"624","volume":"2017","author":"M Schulder","year":"2017","unstructured":"Schulder M, Wiegand M, Ruppenhofer J, Roth B (2017) Towards bootstrapping a polarity shifter lexicon using linguistic features. IJCNLP 2017:624\u2013633","journal-title":"IJCNLP"},{"key":"122_CR4","doi-asserted-by":"crossref","unstructured":"Ito, T, Sakaji, H, Tsubouchi, K, Izumi, K, Yamashita, T (2018) Text-visualizing neural network model: understanding online financial textual data. In: PAKDD 2018","DOI":"10.1007\/978-3-319-93040-4_20"},{"key":"122_CR5","doi-asserted-by":"crossref","unstructured":"Hutto C, Gilbert E (2014) Vader: a parsimonious rule-based model for sentiment analysis of social media text. In :ICWSM-14","DOI":"10.1609\/icwsm.v8i1.14550"},{"key":"122_CR6","unstructured":"Zou QZXHY, Gui T (2018) A lexicon-based supervised attention model for neural sentiment analysis. In: COLING 2018"},{"key":"122_CR7","unstructured":"Quanshi Z, Wu YN, Zhu SC (2018) Interpretable convolutional neural networks. In: CVPR 2018"},{"key":"122_CR8","first-page":"219","volume":"2016","author":"DT Vo","year":"2016","unstructured":"Vo DT, Zhang Y (2016) Don\u2019t count, predict! an automatic approach to learning sentiment lexicons for short text. ACL 2016:219\u2013224","journal-title":"ACL"},{"key":"122_CR9","first-page":"301","volume":"2017","author":"Q Li","year":"2017","unstructured":"Li Q (2017) Learning stock market sentiment lexicon and sentiment-oriented word vector from stocktwits. CoNLL 2017:301\u2013310","journal-title":"CoNLL"},{"issue":"7","key":"122_CR10","first-page":"1","volume":"10","author":"S Bach","year":"2017","unstructured":"Bach S, Binder A, Montavon G, Klauschen F, Muller KR, Samek W (2017) On pixel-wise explanations for nonlinear classifier decisions by layer-wise relevance propagation. PLoS ONE 10(7):1\u201346","journal-title":"PLoS ONE"},{"key":"122_CR11","unstructured":"Sundararajan M, Taly A, Yan Q (2017) Axiomatic attribution for deep networks. In: ICML"},{"key":"122_CR12","unstructured":"Karen S, Andrea V, Zisserman A (2013) Deep inside convolutional networks: visualising image classification models and saliency maps. arXiv:1312.6034"},{"key":"122_CR13","unstructured":"Hechtlinger Y (2016) Interpretation of prediction models using the input gradient. In: arXiv:1611.07634"},{"key":"122_CR14","unstructured":"Springenberg, JT, Dosovitskiy A, Brox T, Riedmiller MA (2015) Striving for simplicity: the all convolutional net. In: ICLR workshop"},{"issue":"11","key":"122_CR15","doi-asserted-by":"publisher","first-page":"2673","DOI":"10.1109\/78.650093","volume":"45","author":"M Schuster","year":"1997","unstructured":"Schuster M, Paliwal K (1997) Bidirectional recurrent neural networks. IEEE Trans Signal Process 45(11):2673\u20132681","journal-title":"IEEE Trans Signal Process"},{"key":"122_CR16","unstructured":"Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, Kaiser Lu, Polosukhin I (2016) Attention is all you need. In: NIPS 2017"},{"key":"122_CR17","doi-asserted-by":"crossref","unstructured":"Wang W, Yang N, Wei F, Chang B, Zhou M (2017) Gated self-matching networks for reading comprehension and question answering. In: ACL 2017","DOI":"10.18653\/v1\/P17-1018"},{"key":"122_CR18","unstructured":"Mikolov T, Sutskever I, Chen K, Corrado G, Dean J (2013) Distributed representations of words and phrases and their compositionality. In: NIPS 2013"},{"issue":"10","key":"122_CR19","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18637\/jss.v050.i10","volume":"50","author":"MKK Hornik","year":"2012","unstructured":"Hornik MKK, Feinerer I, Buchta C (2012) Spherical k-means clustering. J Stat Softw 50(10):1\u201322","journal-title":"J Stat Softw"},{"key":"122_CR20","unstructured":"Nakov P, Rosenthal S, Kozareva, Stoyanov V, Ritter A, Wilson T (2013) Semeval-2013 task 2: sentiment analysis in twitter. In: SemEval 2013"},{"key":"122_CR21","doi-asserted-by":"crossref","unstructured":"Rosenthal S, Nakov P, Ritter A, Stoyanov V (2014) Semeval-2014 task 9: sentiment analysis in twitter. In: SemEval 2014","DOI":"10.3115\/v1\/S14-2009"},{"key":"122_CR22","unstructured":"Zhao P, Zhang T (2014) Accelerating minibatch stochastic gradient descent using stratified sampling.arXiv:1405.3080v1"},{"key":"122_CR23","unstructured":"Kingma JLBDP (2014) Adam: a method for stochastic optimization. arXiv:1412.6980"},{"issue":"1","key":"122_CR24","first-page":"1929","volume":"15","author":"N Srivastava","year":"2014","unstructured":"Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R (2014) Dropout: a simple way to prevent neural networks from overfitting. JMLR 15(1):1929\u20131958","journal-title":"JMLR"},{"key":"122_CR25","unstructured":"Mohammad S, Kiritchenko S, Zhu XD (2013) NRC-Canada: building the state-of-the-art in sentiment analysis of tweets. In: SemEval-2013"},{"key":"122_CR26","doi-asserted-by":"crossref","unstructured":"Fancellu F, Lopez A, Webber B (2016) Neural networks for negation scope detection. In: ACL 2016","DOI":"10.18653\/v1\/P16-1047"},{"key":"122_CR27","first-page":"635","volume":"2010","author":"S Li","year":"2010","unstructured":"Li S, Yat S, Lee M, Chen Y, Huang CR, Wang G (2010) Sentiment classification and polarity shifting. COLING 2010:635\u2013643","journal-title":"COLING"},{"key":"122_CR28","doi-asserted-by":"crossref","unstructured":"Yang Z, Yang D, Dyer C, He X, Smola A, Hovy E (2016) Hierarchical attention networks for document classification. In: NAACL 2016","DOI":"10.18653\/v1\/N16-1174"},{"key":"122_CR29","doi-asserted-by":"crossref","unstructured":"Hu Q, Zhou J, Chen Q, He L (2018) SNNN: promoting word sentiment and negation in neural sentiment classification. In: AAAI 2018","DOI":"10.1609\/aaai.v32i1.11665"},{"key":"122_CR30","doi-asserted-by":"crossref","unstructured":"Arras L, Montavon G, Muller KR, Samek W (2017) Explaining recurrent neural network predictions in sentiment analysis. In: EMNLP workshop","DOI":"10.18653\/v1\/W17-5221"},{"key":"122_CR31","doi-asserted-by":"crossref","unstructured":"Ribeiro MT, Singh S, Guestrin C (2016) why should i trust you? Explaining the predictions of any classifier. In: KDD 2016","DOI":"10.18653\/v1\/N16-3020"},{"key":"122_CR32","unstructured":"Shrikumar A, Greenside P, Kundaje A (2017) Learning important features through propagating activation differences. In: ICML"},{"key":"122_CR33","first-page":"347","volume":"2005","author":"T Wilson","year":"2005","unstructured":"Wilson T, Wiebe J, Hoffman P (2005) Recognizing contextual polarity in phrase level sentiment analysis. EMNLP 2005:347\u2013354","journal-title":"EMNLP"},{"key":"122_CR34","first-page":"43","volume":"2016","author":"S Kiritchenko","year":"2016","unstructured":"Kiritchenko S, Mohammad SM (2016) The effect of negators, modals, and degree adverbs on sentiment composition. NAACL-HLT 2016:43\u201352","journal-title":"NAACL-HLT"},{"key":"122_CR35","first-page":"121","volume":"2016","author":"K Labille","year":"2016","unstructured":"Labille K, Alfarhood S, Gauch S (2016) Estimating sentiment via probability and information theory. KDIR 2016:121\u2013129","journal-title":"KDIR"}],"container-title":["Data Science and Engineering"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s41019-020-00122-4.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s41019-020-00122-4\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s41019-020-00122-4.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,10,1]],"date-time":"2023-10-01T09:22:21Z","timestamp":1696152141000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s41019-020-00122-4"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,5,20]]},"references-count":35,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2020,6]]}},"alternative-id":["122"],"URL":"https:\/\/doi.org\/10.1007\/s41019-020-00122-4","relation":{},"ISSN":["2364-1185","2364-1541"],"issn-type":[{"value":"2364-1185","type":"print"},{"value":"2364-1541","type":"electronic"}],"subject":[],"published":{"date-parts":[[2020,5,20]]},"assertion":[{"value":"12 March 2020","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"28 April 2020","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"30 April 2020","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"20 May 2020","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}}]}}