{"title":"Effective Candidate Selection and Interpretable Interest Extraction for Follower Prediction on Social Media","authors":"Seiji Maekawa, Santi Saeyor, Takeshi Sakaki, Makoto Onizuka","doi":"10.1145/3486622.3493954","DOIUrl":null,"url":null,"abstract":"We address the problem of predicting new followers for a company account given that the number of social network API calls is limited, in order to enhance the marketing communication effectiveness on social media. The contributions of this paper are three-fold: 1) filtering methods that select promising candidate accounts with high precision while effectively reducing the number of API calls, 2) a new method for extracting interpretable feature vectors (interest vectors) from each account by utilizing standardized categories for marketing communication, and 3) a follower prediction model by utilizing the above candidate selection methods and interpretable interest vectors. Experiments on Twitter data confirm that our follower prediction model performs well with a small number of API calls and clarify which dimension of interest vectors (interest category) contributes to prediction performance.","PeriodicalId":89230,"journal":{"name":"Proceedings. IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology","volume":"73 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2021-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings. IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3486622.3493954","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
We address the problem of predicting new followers for a company account given that the number of social network API calls is limited, in order to enhance the marketing communication effectiveness on social media. The contributions of this paper are three-fold: 1) filtering methods that select promising candidate accounts with high precision while effectively reducing the number of API calls, 2) a new method for extracting interpretable feature vectors (interest vectors) from each account by utilizing standardized categories for marketing communication, and 3) a follower prediction model by utilizing the above candidate selection methods and interpretable interest vectors. Experiments on Twitter data confirm that our follower prediction model performs well with a small number of API calls and clarify which dimension of interest vectors (interest category) contributes to prediction performance.