[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82316-en":3,"doc-seo-82316-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},82316,1374391974564,"Clementine","https://ap-avatar.wpscdn.com/avatar/14000253aa45c000a9e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779874745381141002",8,"Research & Report","DKCD Domain Knowledge-Enhanced Causal Discovery from Unstructured Data","Causal discovery from unstructured data is an underexplored yet challenging task in high-expertise areas such as healthcare, finance, and education. Existing approaches often use large language models to extract causal factors and convert them into structured inputs for causal graph construction, but they are constrained by two issues: incomplete identification of latent factors and unreliable factor annotation due to insufficient domain-grounded reasoning. DKCD introduces knowledge mining, knowledge-guided causal reasoning, and causal structure discovery. Experiments on two domain-specific datasets show improvements in both factor identification and causal graph construction.","DKCD: Domain Knowledge-Enhanced Causal Discovery from  \nUnstructured Data  \nXin Li and Jin Li and Shoujin Wang and Kun Yu and Fang Chen  \nUniversity of Technology Sydney  \nSydney, NSW, Australia  \n[xin.li-19@student.uts.edu.au](xin.li-19@student.uts.edu.au)  \n[jin.li-4@student.uts.edu.au](jin.li-4@student.uts.edu.au)  \n[shoujin.wang@uts.edu.au](shoujin.wang@uts.edu.au)  \n[Kun.Yu@uts.edu.au](Kun.Yu@uts.edu.au)  \n[Fang.Chen@uts.edu.au](Fang.Chen@uts.edu.au)  \narXiv :2607 .09348v 1 [ cs .CL] 10 Jul 2026  \nAbstract  \nCausal discovery from unstructured data is a challenging yet underexplored task in highexpertise domains such as healthcare, finance, and education. Existing methods typically leverage the general knowledge of large language models (LLMs) to identify causal factors from unstructured data and annotate them into structured data for causal graph construction. However, they remain limited by two key challenges (CHs): (CH1) insufficient identification of latent factors, which are implicit in the data yet essential for causal discovery, due to the lack of domain-specific knowledge; and (CH2) unreliable factor annotation, caused by the lack of domain-grounded reasoning, which propagates errors to the resulting causal graphs. To address these challenges, we introduce a novel Domain Knowledge-enhanced Causal Discovery framework (DKCD) for causal discovery from unstructured data in highexpertise domains with three interconnected components: (1) Knowledge Mining: It retrieves relevant domain knowledge based on observable factors to support subsequent causal reasoning. (2) Knowledge-guided Causal Reasoning: Reasoning with relevant knowledge, it discovers latent causal factors to address CH1 and generates key causal clues for more accurate data annotation to address CH2 .  \nand (3) Causal Structure Discovery: It constructs the final causal graphs based on a more complete factor set and accurate annotations. Experiments on two domain-specific datasets show that DKCD significantly improves both causal factor identification and causal graph construction.  \n1 Introduction  \nExploring and uncovering causality from realworld data is fundamental across a wide range of scientific domains (Bunge, 2017 ; Illari et al., 2011), including healthcare (Shi and Norgeot, 2022 ; Yang et al., 2013), finance (Papana et al., 2017 ; Sokolov  \net al., 2025), and education (Morrison and van der Werf, 2016 ; Fancsali, 2014) . In data-driven causal discovery, causal relationships are typically represented as causal graphs. Existing causal discovery algorithms, such as the Peter–Clark (PC) algorithm (Spirtes et al., 2000 ; Ramsey, 2016) and the Fast Causal Inference (FCI) algorithm (Spirtes et al., 2013), are primarily designed for structured data with predefined schemes and fixed formats, commonly organized in tabular form (e.g., relational databases and spreadsheets) . Thus, they face substantial limitations when applied to unstructured data (Malinsky and Danks, 2018 ; Yu et al., 2016) . Meanwhile, the growing dominance of unstructured data, particularly natural-language text (Siddiqa et al., 2017 ; Eberendu et al., 2016 ; Azad et al., 2020), calls for methods that bridge unstructured data and statistical causal discovery.  \nWith the emergence of large language models (LLMs), new opportunities have arisen for analyzing and discovering causal relationships from unstructured data (Wang et al., 2023 ; Wang, 2024) . With strong capabilities for understanding, reasoning, and extracting knowledge from natural language, LLMs can support automatic identification of causal factors, reducing reliance on manual domain expertise for data labeling (Ashwani et al., 2024 ; Dubey et al., 2024) . Specifically, the COAT framework (Liu et al., 2024) is a representative approach that leverages LLMs to extract causal factors from unstructured data, annotate factor values, and then apply statistical causal discovery algorithms to construct causal graphs. Although COAT takes an","cbCainWk1RQ9sPWo","https://ap.wps.com/l/cbCainWk1RQ9sPWo","pdf",1908110,1,25,"English","en",105,"# Introduction\n## Causal discovery from unstructured data\n## Using LLMs for causal factor extraction\n## Challenges in high-expertise domains","[{\"question\":\"What are the two main challenges (CH1 and CH2) DKCD targets in causal discovery from unstructured data?\",\"answer\":\"CH1 is insufficient identification of latent factors because domain-specific knowledge is missing. CH2 is unreliable factor annotation due to lack of domain-grounded reasoning, which then propagates errors into the final causal graphs.\"},{\"question\":\"How does DKCD incorporate domain knowledge to improve causal factor identification?\",\"answer\":\"DKCD uses a Knowledge Mining component to retrieve relevant domain knowledge based on observable factors. This knowledge supports knowledge-guided causal reasoning to discover latent causal factors that are implicit in the unstructured data.\"},{\"question\":\"What is DKCD’s overall pipeline for constructing causal graphs?\",\"answer\":\"DKCD first mines domain knowledge, then performs knowledge-guided causal reasoning to generate key causal clues and more accurate annotations. Finally, it performs causal structure discovery to construct causal graphs using a more complete factor set and improved annotations.\"}]",1784179565,63,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"dkcd-domain-knowledge-enhanced-causal-discovery-from-unstructured-data","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/dkcd-domain-knowledge-enhanced-causal-discovery-from-unstructured-data/82316/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What are the two main challenges (CH1 and CH2) DKCD targets in causal discovery from unstructured data?","Question",{"text":75,"@type":76},"CH1 is insufficient identification of latent factors because domain-specific knowledge is missing. CH2 is unreliable factor annotation due to lack of domain-grounded reasoning, which then propagates errors into the final causal graphs.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does DKCD incorporate domain knowledge to improve causal factor identification?",{"text":80,"@type":76},"DKCD uses a Knowledge Mining component to retrieve relevant domain knowledge based on observable factors. This knowledge supports knowledge-guided causal reasoning to discover latent causal factors that are implicit in the unstructured data.",{"name":82,"@type":73,"acceptedAnswer":83},"What is DKCD’s overall pipeline for constructing causal graphs?",{"text":84,"@type":76},"DKCD first mines domain knowledge, then performs knowledge-guided causal reasoning to generate key causal clues and more accurate annotations. Finally, it performs causal structure discovery to construct causal graphs using a more complete factor set and improved annotations.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]