Jorge Alvarez-Lozano, J. A. García-Macías, Edgar Chávez
{"title":"位置预测中的学习与用户自适应","authors":"Jorge Alvarez-Lozano, J. A. García-Macías, Edgar Chávez","doi":"10.1145/2494091.2495978","DOIUrl":null,"url":null,"abstract":"User location forecasting is central to establish context in proactive mobile applications. Knowing where the user will be at a given time enables standby action triggers ahead of time. User location exhibits periodic patterns grouped by time of day, day of the week, month of the year, etc. This characteristic has been exploited to model user location as a Markov process with great accuracy. Using yearly data from public sources it was possible to predict user location in a time frame of 8 hours with accuracy of up to 69%. One assumption of the above modeling is that user location is stationary in time. However, it is more natural to assume user location patterns may vary over time. For example one user may change job, or the relationship status, and avoid certain places frecuented in the past. In this paper we propose a learning mechanism adapting user location forecasting to behavior changes over time. Our model is able to predict for up to 94 weeks with 43% of accuracy.","PeriodicalId":220524,"journal":{"name":"Proceedings of the 2013 ACM conference on Pervasive and ubiquitous computing adjunct publication","volume":"3 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2013-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"19","resultStr":"{\"title\":\"Learning and user adaptation in location forecasting\",\"authors\":\"Jorge Alvarez-Lozano, J. A. García-Macías, Edgar Chávez\",\"doi\":\"10.1145/2494091.2495978\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"User location forecasting is central to establish context in proactive mobile applications. Knowing where the user will be at a given time enables standby action triggers ahead of time. User location exhibits periodic patterns grouped by time of day, day of the week, month of the year, etc. This characteristic has been exploited to model user location as a Markov process with great accuracy. Using yearly data from public sources it was possible to predict user location in a time frame of 8 hours with accuracy of up to 69%. One assumption of the above modeling is that user location is stationary in time. However, it is more natural to assume user location patterns may vary over time. For example one user may change job, or the relationship status, and avoid certain places frecuented in the past. In this paper we propose a learning mechanism adapting user location forecasting to behavior changes over time. Our model is able to predict for up to 94 weeks with 43% of accuracy.\",\"PeriodicalId\":220524,\"journal\":{\"name\":\"Proceedings of the 2013 ACM conference on Pervasive and ubiquitous computing adjunct publication\",\"volume\":\"3 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2013-09-08\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"19\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the 2013 ACM conference on Pervasive and ubiquitous computing adjunct publication\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/2494091.2495978\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 2013 ACM conference on Pervasive and ubiquitous computing adjunct publication","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/2494091.2495978","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Learning and user adaptation in location forecasting
User location forecasting is central to establish context in proactive mobile applications. Knowing where the user will be at a given time enables standby action triggers ahead of time. User location exhibits periodic patterns grouped by time of day, day of the week, month of the year, etc. This characteristic has been exploited to model user location as a Markov process with great accuracy. Using yearly data from public sources it was possible to predict user location in a time frame of 8 hours with accuracy of up to 69%. One assumption of the above modeling is that user location is stationary in time. However, it is more natural to assume user location patterns may vary over time. For example one user may change job, or the relationship status, and avoid certain places frecuented in the past. In this paper we propose a learning mechanism adapting user location forecasting to behavior changes over time. Our model is able to predict for up to 94 weeks with 43% of accuracy.