{"corpus_id":245206961,"paper_sha":"a481f69879c9acd3749d395c797fbfe3f197cc9f","doi":"10.1155/2021/6913427","arxiv_id":null,"pmid":null,"pmcid":null,"mag_id":null,"dblp_id":"journals/sp/HuangW21a","acl_id":null,"title":"Research on Deep Learning-Based Financial Risk Prediction","year":2021,"publication_date":"2021-12-13","venue":"Scientific Programming","journal":{"name":"Sci. Program.","pages":"6913427:1-6913427:8","volume":"2021"},"journal_issn":null,"journal_title":null,"publication_types":["JournalArticle"],"pubmed_pub_types":null,"s2_fields_of_study":["Business","Computer Science"],"reference_count":32,"citation_count":7,"influential_citation_count":0,"is_open_access":true,"arxiv_categories":null,"arxiv_license":null,"arxiv_journal_ref":null,"mesh_headings":null,"chemicals":null,"comments_corrections":null,"source_flags":1,"s2_open_access_pdf_url":"https://downloads.hindawi.com/journals/sp/2021/6913427.pdf","s2_open_access_landing_url":"https://www.semanticscholar.org/paper/a481f69879c9acd3749d395c797fbfe3f197cc9f","s2_open_access_license":null,"s2_open_access_status":"GOLD","pmc_open_access_pdf_url":null,"pmc_open_access_landing_url":null,"pmc_open_access_license":null,"pmc_open_access_status":null,"unpaywall_open_access_pdf_url":null,"unpaywall_open_access_landing_url":null,"unpaywall_open_access_license":null,"unpaywall_open_access_status":null,"abstract":"Financial text-based risk prediction is an important subset for financial analysis. Through automatic analysis of public financial comments, fundamentals on current financial expectations can be evaluated. A deep learning method for financial risk prediction based on sentiment classification is proposed in this paper. The proposed method consists of two steps. Firstly, the abstract of the financial message is extracted according to the seq2seq model. During the extraction process, the seq2seq model can cope with the situation of different input message lengths. After the abstraction, invalid information in the financial messages can be effectively filtered, thus accelerating the subsequent sentiment classification step. The sentiment classification step is performed through the GRU model according to the abstracted texts. The proposed method has the following advantages: (1) it can handle financial messages of different lengths; (2) it can filter out the invalid information of financial messages; (3) because the extracted abstract is more refined, it can speed up the subsequent sentiment classification step; and (4) it has better sentiment classification accuracy. The proposed method in this paper is then verified through financial message dataset from the financial social network StockTwits. By comparing the classification performances, it can be seen that compared with the classical SVM and LSTM methods, the proposed method in this paper can improve the accuracy of sentiment classification by 5.57% and 2.58%, respectively.","claims":[{"public_id":"cl_58d73429d47b6aec2e436cc8cd3b729d","status":"active","text":"A seq2seq-based abstraction step followed by GRU sentiment classification is proposed for financial risk prediction.","confidence":0.98,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_58d73429d47b6aec2e436cc8cd3b729d"},{"public_id":"cl_764c6bdddcd9709d206e877acaa13b88","status":"active","text":"On the StockTwits financial message dataset, the proposed method improves sentiment classification accuracy by 5.57% over SVM and 2.58% over LSTM.","confidence":0.99,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_764c6bdddcd9709d206e877acaa13b88"},{"public_id":"cl_b423906a38544df1e8a5bc0397cf65e8","status":"active","text":"The abstraction step handles financial messages of different lengths and filters invalid information before sentiment classification.","confidence":0.93,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_b423906a38544df1e8a5bc0397cf65e8"},{"public_id":"cl_46a8d5516f28aa8a3509e1855dd94387","status":"active","text":"Using the extracted abstract speeds up the subsequent sentiment classification step.","confidence":0.88,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_46a8d5516f28aa8a3509e1855dd94387"}],"concepts":[{"public_id":"co_02472aed7b78df4b08f091df2fe4addb","status":"active","name":"sentiment classification","description":"The classification of text into sentiment categories for financial risk assessment.","types":["task"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_02472aed7b78df4b08f091df2fe4addb"},{"public_id":"co_1a1f6677990423d046a8b8455832b2bb","status":"active","name":"seq2seq model","description":"A sequence-to-sequence neural model used here to extract an abstract from financial messages.","types":["model"],"aliases":["seq2seq"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_1a1f6677990423d046a8b8455832b2bb"},{"public_id":"co_298d0d994beb8513facf60c6517dc686","status":"active","name":"financial risk prediction","description":"The task of estimating financial risk from text-based signals and related financial information.","types":["task"],"aliases":["financial text-based risk prediction"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_298d0d994beb8513facf60c6517dc686"},{"public_id":"co_2c0eee6ac923d3cf19187a5386254ace","status":"active","name":"SVM","description":"A support vector machine baseline used for classification comparison.","types":["baseline","classifier"],"aliases":["support vector machine"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_2c0eee6ac923d3cf19187a5386254ace"},{"public_id":"co_34c30b542cc0c7a0a4c23e965928542b","status":"active","name":"abstracted texts","description":"Refined text representations produced after extracting abstracts from the original financial messages.","types":["text representation"],"aliases":["abstract"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_34c30b542cc0c7a0a4c23e965928542b"},{"public_id":"co_357ee1cb2dd54e650bfc3018b81f3d07","status":"active","name":"financial messages","description":"Textual financial posts or comments used as input data for analysis.","types":["text data"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_357ee1cb2dd54e650bfc3018b81f3d07"},{"public_id":"co_89db2d4bd0f07ee079bab71ecc0686f2","status":"active","name":"GRU model","description":"A gated recurrent unit neural network used here for sentiment classification.","types":["model"],"aliases":["GRU"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_89db2d4bd0f07ee079bab71ecc0686f2"},{"public_id":"co_8d89656b4a8ced2d1d0d0576e49bc778","status":"active","name":"StockTwits","description":"A financial social network used as the source of the evaluation dataset.","types":["dataset source","social network"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_8d89656b4a8ced2d1d0d0576e49bc778"},{"public_id":"co_a850dc1b7610aee12f102d9bccfc5fb7","status":"active","name":"LSTM","description":"A long short-term memory baseline used for classification comparison.","types":["baseline","classifier"],"aliases":["long short-term memory"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_a850dc1b7610aee12f102d9bccfc5fb7"}],"external_ids":{"DOI":"10.1155/2021/6913427","ArXiv":null,"PubMed":null,"PubMedCentral":null,"MAG":null,"DBLP":"journals/sp/HuangW21a","ACL":null},"open_access":{"is_open_access":true,"pdf_url":"https://downloads.hindawi.com/journals/sp/2021/6913427.pdf","landing_url":"https://www.semanticscholar.org/paper/a481f69879c9acd3749d395c797fbfe3f197cc9f","source":"semantic_scholar","pdf_url_source":"semantic_scholar_open_access_pdf","license":null,"status":"GOLD","reason":null},"reference_availability":{"status":"available","references_indexed":true,"full_text_available":false,"full_text_source":null,"count_basis":"semantic_scholar_metadata","extraction_status":"not_applicable","reason":null},"source":{"provider":"episteme2","base_corpus":"semantic_scholar_dump","freshness_mode":"unknown","basis":["semantic_scholar_metadata","postgres_metadata"],"limits":["paper metadata is based on indexed upstream scholarly datasets","claims and concepts are available only for extracted papers","absence of claims or concepts means no extracted graph data is available in this response"],"status":"available","degraded":false,"degraded_reasons":[],"diagnostics":{"status":"available","degraded":false,"degraded_reasons":[],"metadata_status":"available","graph_status":"available","abstract_status":"available"},"source_flags":1},"paper_id":634146,"paper_uid":"6334d060-bf02-4fdd-87c2-f0ae75f933d8","canonical_identity":{"paper_id":634146,"paper_uid":"6334d060-bf02-4fdd-87c2-f0ae75f933d8","identity_status":"available","lookup_basis":"semantic_scholar_external_id","compatibility_path":"corpus_id"},"url":"https://sah.borca.ai/papers/245206961"}