[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-1-en-105":3,"doc-seo-189689-105":53,"doc-detail-189689-en":126},{"code":4,"msg":5,"data":6},0,"success",[7,14,19,24,29,34,39,44,49],{"id":8,"doc_module":9,"doc_module_name":10,"category_name":11,"show_sort_weight":12,"slug":13},11,1,"Template","Presentations",90,"presentations",{"id":15,"doc_module":9,"doc_module_name":10,"category_name":16,"show_sort_weight":17,"slug":18},12,"Resumes",80,"resumes",{"id":20,"doc_module":9,"doc_module_name":10,"category_name":21,"show_sort_weight":22,"slug":23},14,"Invoices",70,"invoices",{"id":25,"doc_module":9,"doc_module_name":10,"category_name":26,"show_sort_weight":27,"slug":28},15,"Posters",60,"posters",{"id":30,"doc_module":9,"doc_module_name":10,"category_name":31,"show_sort_weight":32,"slug":33},16,"Social Media",50,"social-media",{"id":35,"doc_module":9,"doc_module_name":10,"category_name":36,"show_sort_weight":37,"slug":38},17,"Forms",40,"forms",{"id":40,"doc_module":9,"doc_module_name":10,"category_name":41,"show_sort_weight":42,"slug":43},18,"Letters",30,"letters",{"id":45,"doc_module":9,"doc_module_name":10,"category_name":46,"show_sort_weight":47,"slug":48},21,"Paper Templates",5,"papers-templates",{"id":50,"doc_module":9,"doc_module_name":10,"category_name":51,"show_sort_weight":4,"slug":52},158,"General","general-158",{"code":4,"msg":54,"data":55},"ok",{"site_id":56,"language":57,"slug":58,"title":59,"keywords":60,"description":61,"schema_data":62,"social_meta":119,"head_meta":121,"extra_data":123,"updated_unix":125},105,"en","document-metadata-extraction-multimodal","Document Metadata Extraction (Multimodal)","","This document appears to be a technical report or guide detailing methods for document metadata extraction, with a particular focus on multimodal analysis that integrates both text and image data. It outlines various similarity metrics used in processing, such as Cosine Similarity and KL Divergence, and discusses different edge types (Undirected, Directed). The report also presents a classification of email types (Finance, Receipt, Travel), including their respective template counts and performance metrics across different labeling and similarity approaches like Majority Label, Centroid Similarity, and Label Propagation. Furthermore, it delves into topic modeling and Bag-of-Words techniques, providing precision and recall scores for various methods. The document concludes with a summary of email coverage based on different categories and a breakdown of unlabeled topics ranging from Business & Tech to Politics and Donations.",{"@graph":63,"@context":118},[64,80,101],{"@type":65,"itemListElement":66},"BreadcrumbList",[67,71,74,77],{"item":68,"name":69,"@type":70,"position":9},"https://docshare.wps.com","Home","ListItem",{"item":72,"name":10,"@type":70,"position":73},"https://docshare.wps.com/template/",2,{"item":75,"name":51,"@type":70,"position":76},"https://docshare.wps.com/template/general/",3,{"item":78,"name":59,"@type":70,"position":79},"https://docshare.wps.com/template/document-metadata-extraction-multimodal/189689/",4,{"url":78,"name":59,"@type":81,"image":82,"author":87,"headline":59,"publisher":90,"fileFormat":93,"inLanguage":57,"description":61,"dateModified":94,"datePublished":95,"encodingFormat":93,"isAccessibleForFree":96,"interactionStatistic":97},"DigitalDocument",{"url":83,"@type":84,"width":85,"height":86},"https://docshare.wps.com/thumbnails/document-metadata-extraction-multimodal/189689.png","ImageObject",442,249,{"name":88,"@type":89},"Seraphina","Person",{"url":68,"name":91,"@type":92},"DocShare","Organization","application/pdf","2026-10-02","2026-09-03",true,{"@type":98,"interactionType":99,"userInteractionCount":79},"InteractionCounter",{"@type":100},"ViewAction",{"@type":102,"mainEntity":103},"FAQPage",[104,110,114],{"name":105,"@type":106,"acceptedAnswer":107},"What are the primary methods discussed for document similarity?","Question",{"text":108,"@type":109},"The document mentions Cosine Similarity and KL Divergence as key methods for calculating similarity between documents, alongside specific edge types like Undirected and Directed.","Answer",{"name":111,"@type":106,"acceptedAnswer":112},"What types of emails are classified and analyzed in the document?",{"text":113,"@type":109},"The document classifies emails into Finance, Receipt, and Travel categories, providing metrics on labeled templates and performance across various analytical approaches.",{"name":115,"@type":106,"acceptedAnswer":116},"What are the different approaches presented for document classification and analysis?",{"text":117,"@type":109},"The document outlines several approaches, including Majority Label, Centroid Similarity, Label Propagation, and Bag-of-Words (BoW) with both KL Divergence (KL) and Cosine Similarity (CS) metrics.","https://schema.org",{"og:url":78,"og:type":120,"og:title":59,"og:site_name":91,"og:description":61},"article",{"robots":122,"canonical":78},"index,follow",{"doc_id":124,"site_id":56},189689,1788398565,{"code":4,"msg":5,"data":127},{"doc_id":124,"user_id":128,"nickname":88,"user_avatar":129,"doc_module":9,"category_id":50,"category_name":51,"doc_title":59,"doc_description":61,"doc_content":130,"file_id":131,"file_url":132,"file_type":133,"file_size":134,"view_count":135,"is_deleted":4,"is_public":9,"is_downloadable":9,"audit_status":9,"page_count":136,"language":137,"language_code":57,"site_id":56,"html_lang":57,"table_of_contents":138,"faqs":139,"seo_title":140,"seo_description":61,"update_tm":125,"read_time":79},962075114101,"https://ap-avatar.wpscdn.com/avatar/e000253a75eb197efd?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780044092746381165","| Method | Similarity Metric | Edge Type |\n| --- | --- | --- |\n| Cosine Similarity\u003Cbr>KL Divergence | Px2FTi ;FTj w (x;Ti)w (x;Tj)\u003Cbr>qPx2FTi w (x;Ti)2 qPx2FTj w (x;Tj)2 exp(􀀀Px2FTi \\FTj p (xj Ti ) log p􀀘p((xxjj~~ ~~TT~~i~~j)) ) | Undirected\u003Cbr>Directed |\n\n\n| Label | Example of email types | \\# Labeled Templates |\n| --- | --- | --- |\n| Finance | Financial statements, stock reports, bank account updates. | 262 |\n| Receipt | Purchase receipts, order con􀀌rmations, shipping notices. | 432 |\n| Travel | Travel itineraries, hotel reservations, car rentals. | 201 |\n\n\n|  | Majority Label | Centroid Similarity | Label Propagation |\n| --- | --- | --- | --- |\n|  | Precision Recall | Precision Recall | Precision Recall |\n| Finance | 55.00 83.97 | 60.73 89.30 | 91.49 98.47 m; c |\n| Receipt | 89.35 79.63 | 85.05 95.87 | 97.82 93.52 c |\n| Travel | 86.57 86.57 | 86.62 95.33 | 95.00 94.53 m; c |\n\n\n|  | Topics [KL] | Topics [CS] | BoW [KL] | BoW [CS] |\n| --- | --- | --- | --- | --- |\n|  | Precision Recall | Precision Recall | Precision Recall | Precision Recall |\n| Finance | 80.44 83.21 | 83.09 86.26 | 93.13 93.13 | 91.49 98.47 |\n| Receipt | 88.40 82.87 | 89.78 85.42 | 93.47 96.06 | 97.82 93.52 |\n| Travel | 88.61 89.05 | 86.79 91.54 | 95.24 89.55 | 95.00 94.53 tcs |\n\n|  | Precision Recall | Email coverage |\n| --- | --- | --- |\n| Finance\u003Cbr>Receipt Travel | 93.08 99.64 99.13 91.39 95.93 93.52 | +9:7%\u003Cbr>+16:8%\u003Cbr>+5:7% |\n\n| Unlabeled |  |  |  |  |  |  |\n| --- | --- | --- | --- | --- | --- | --- |\n| Business & Tech | Music |  | Social | Fashion | Politics |  |\n| develop\u003Cbr>technolog\u003Cbr>job\u003Cbr>learn\u003Cbr>compani | music\u003Cbr>world\u003Cbr>love\u003Cbr>artist\u003Cbr>star | music\u003Cbr>ticket\u003Cbr>show\u003Cbr>live\u003Cbr>band | story\u003Cbr>her\u003Cbr>his\u003Cbr>video\u003Cbr>friend | fashion\u003Cbr>deal\u003Cbr>sale\u003Cbr>design\u003Cbr>brand | say\u003Cbr>govern\u003Cbr>american\u003Cbr>polit\u003Cbr>obama | thank\u003Cbr>contribut\u003Cbr>support\u003Cbr>campaign\u003Cbr>donat |","cbCainMhydbm1juA","https://ap.wps.com/l/cbCainMhydbm1juA","pdf",745304,6,10,"English","# Document Metadata Extraction (Multimodal)\n## Instructions\n## Language Rules\n## Field Specifications\n### Title Rules\n### `title_en`\n### `language`\n### abstract\n### `keywords`\n### `toc`\n### `faqs`\n### `category`\n## Output Requirements (STRICT)\n## Inputs\n**File name:**\n**Document title:**\n**Category definitions:**\n**Document text:**\n**Document images:**","[{\"question\":\"What are the primary methods discussed for document similarity?\",\"answer\":\"The document mentions Cosine Similarity and KL Divergence as key methods for calculating similarity between documents, alongside specific edge types like Undirected and Directed.\"},{\"question\":\"What types of emails are classified and analyzed in the document?\",\"answer\":\"The document classifies emails into Finance, Receipt, and Travel categories, providing metrics on labeled templates and performance across various analytical approaches.\"},{\"question\":\"What are the different approaches presented for document classification and analysis?\",\"answer\":\"The document outlines several approaches, including Majority Label, Centroid Similarity, Label Propagation, and Bag-of-Words (BoW) with both KL Divergence (KL) and Cosine Similarity (CS) metrics.\"}]","Document Metadata Extraction (Multimodal) | PDF"]