[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-1-en-105":3,"doc-seo-195224-105":53,"doc-detail-195224-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","verbalizing-implicit-preferences-vip-framework-algorithm-1-rating-prediction-and-recommendation","Verbalizing Implicit Preferences (VIP) Framework - Algorithm 1 - Rating Prediction and Recommendation","","VIP (Verbalizing Implicit Preferences) framework converts a user’s review history and item metadata into an explicit preference representation, then predicts ratings and produces ranked recommendations. The method first generates a user profile via review analysis and review fusion using a large language model with tailored prompting strategies. It also generates a product profile from item metadata and computes rating prediction by scoring the match between user and product profiles, returning either a single predicted rating or a recommendation list.",{"@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/verbalizing-implicit-preferences-vip-framework-algorithm-1-rating-prediction-and-recommendation/195224/",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/verbalizing-implicit-preferences-vip-framework-algorithm-1-rating-prediction-and-recommendation/195224.png","ImageObject",442,249,{"name":88,"@type":89},"wps_ap_test_251126_0180","Person",{"url":68,"name":91,"@type":92},"DocShare","Organization","application/pdf","2026-09-26","2026-09-03",true,{"@type":98,"interactionType":99,"userInteractionCount":47},"InteractionCounter",{"@type":100},"ViewAction",{"@type":102,"mainEntity":103},"FAQPage",[104,110,114],{"name":105,"@type":106,"acceptedAnswer":107},"What does the VIP framework take as input and what does it output?","Question",{"text":108,"@type":109},"It takes user review history, item metadata, and a large language model with prompting strategies, and outputs a predicted rating or a recommendation list.","Answer",{"name":111,"@type":106,"acceptedAnswer":112},"How is the user profile generated in the VIP method?",{"text":113,"@type":109},"The framework analyzes each review to extract aspects the user likes or dislikes, then fuses the analysis into an updated user profile representation.",{"name":115,"@type":106,"acceptedAnswer":116},"How are ratings and recommendations produced after user and product profiles are built?",{"text":117,"@type":109},"It generates a product profile from item metadata, then uses the language model to compute a match score between the user profile and product profile, producing a rating prediction or ranked recommendations.","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},195224,1788446475,{"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":47,"is_deleted":4,"is_public":9,"is_downloadable":9,"audit_status":9,"page_count":135,"language":136,"language_code":57,"site_id":56,"html_lang":57,"table_of_contents":137,"faqs":138,"seo_title":139,"seo_description":61,"update_tm":125,"read_time":79},8796095027276,"https://avatar.qwps.com/avatar/d3BzX2FwX3Rlc3RfMjUxMTI2XzAxODA=","| Algorithm 1: Verbalizing Implicit Preferences (VIP) Framework |\n| --- |\n| Input: User review history Du = { (ri , ti)}1 ; Item i metadata mi ;\u003Cbr>Large Language Model L with prompting strategies π\u003Cbr>Output: Predicted rating rˆu,i or\u003Cbr>Recommendation list rˆu,dial\u003Cbr>1. User Profile Generation\u003Cbr>u′ ← ϕ foreach (ri , ti) ∈ Du do\u003Cbr>\u003Cbr>pi ← L 􀀀ri ;πreview 􀀁 // review analysis\u003Cbr>u ← L􀀀 Fagg (u ∪ {pi });πfusion 􀀁 // review fusion\u003Cbr>2. Product Profile Generation vi ← L 􀀀mi ;πprod 􀀁\u003Cbr>3. Recommendation (Rating Prediction) rˆu,i ← L 􀀀 P (u, vi );πrate 􀀁 ;\u003Cbr>return rˆu,i or rˆu,dial |\n\n\n| Prompt π | Example prompt excerpt |\n| --- | --- |\n| System Prompt | You are a helpful assistant designed to analyze user preferences and recommend {{Product Category}} . |\n| Review Analysis (πreview ) -Sorted Preferences | You will be given a user profile and a user review for a restaurant. Based on the given information, please answer the following three questions: (1) Record History Ratings: -Extract and list the history ratings given by the user. Format the list as [{{Product Category}} Name-Rating, ...] . (2) Determine User Preferences:\u003Cbr>-Analyze the user review to identify what aspects of the {{Product Category}} the user likes and dislikes. Clearly separate likes from dislikes in your response. (3) Sort User Values: -Based on the user’s historical ratings and the identified likes and dislikes, sort the aspects that will affect the user according to how the user values them. This can include aspects such as {{Domain-specific Attributes}}, etc. When answering, please follow the order of the questions. After you finish the above three questions, please update the history rating list, combine the analyzed points with the original user profile, and shorten the results to make a new profile. |\n| Review Analysis (πreview ) -Predefined Preferences | ... (1) {{Predefined Preference 1}} (2) {{Predefined Preference 2}} (3) History ratings of the user (4) other preferences not included in (1) ∼ (3) ... |\n| Review Analysis (πreview ) -Automatic Preference Selection | ... (1) History ratings of the user (2) The dream restaurant the user might prefer... |\n| Review Fusion (πfusion ) | Please extract the user history ratings and the original user profile from your analysis to make a new user profile. |\n\n\n| Prompt π | Example prompt excerpt |\n| --- | --- |\n| System Prompt | You are a helpful assistant designed to analyze user preferences and recommend {{Product Category}} . |\n| Product Profile Template (πprod) | {{Product Name}} is a {{Product Category}} located at {{Address}} . It has a rating of {{Rating}} stars out of 5 . It is open on {{Open Hour}} . The restaurant accepts {{Services}} . |\n| Rating Prediction (πrate ) | You will be given a {{Product Category}} profile and a user profile with the user’s rating history, please analyze whether the {{Product Category}} matches the user’s preferences. The user profile contains the user’s rating history and preferences sorted by priority from high to low. The {{Product Category}} profile contains the {{Product Category}} features. Considering the user rating tendencies, give the {{Product Category}} a score from 1 ∼ 5 representing how much the user will like it. Note that you should think carefully and make short explanation about which feature(s) in restaurant profile meet or not meet user’s preferences. Show the scores at the end of your answer in the following JSON format: { “score”: predicted score for the {{Product Category}} } |\n\n| Method | Strategy | Backbone Model | Search | Tool | Item Metadata | RMSE | MAE |\n| --- | --- | --- | --- | --- | --- | --- | --- |\n| MF\u003Cbr>FM | –\u003Cbr>– | –\u003Cbr>– | ✗✗ |  | ✗✗ | 1.4095\u003Cbr>1.3286 | 1.1521\u003Cbr>1.0372 |\n| LLMRec | Simple Prompting | Llama3.3-70BGPT-4o | ✗✗ |  | ✗✗ | 1.5275\u003Cbr>1.3914 | 0.9918\u003Cbr>0.9640 |\n| RecMind | Self-Inspire Planning | GPT-4o | ✓ |  | ✓ | 1.2485 | 0.8570 |\n| VIP | PP | Llama3.3-70B\u003Cbr>Llama3.3-70B + GPT-4o | ✗✗ |  | ✓✓ | 1.6078\u003Cbr>1.4175 |","cbCaivVsNl9LWzq1","https://ap.wps.com/l/cbCaivVsNl9LWzq1","pdf",877655,10,"English","# Algorithm 1: Verbalizing Implicit Preferences (VIP) Framework\n## Input and Output\n## User Profile Generation\n## Product Profile Generation\n## Recommendation (Rating Prediction)\n# Prompting Strategy π\n## System Prompt\n## Review Analysis (πreview)\n## Review Fusion (πfusion)\n# Methods and Evaluation Results","[{\"question\":\"What does the VIP framework take as input and what does it output?\",\"answer\":\"It takes user review history, item metadata, and a large language model with prompting strategies, and outputs a predicted rating or a recommendation list.\"},{\"question\":\"How is the user profile generated in the VIP method?\",\"answer\":\"The framework analyzes each review to extract aspects the user likes or dislikes, then fuses the analysis into an updated user profile representation.\"},{\"question\":\"How are ratings and recommendations produced after user and product profiles are built?\",\"answer\":\"It generates a product profile from item metadata, then uses the language model to compute a match score between the user profile and product profile, producing a rating prediction or ranked recommendations.\"}]","Verbalizing Implicit Preferences (VIP) Framework - Algorithm 1 - Rating Prediction and Recommendation | PDF"]