24小时热门版块排行榜

返回列表

【悬赏金币】回答本帖问题，作者blairwu将赠送您 10 个金币

blairwu

铜虫 (小有名气)

应助: 3 (幼儿园)
金币: 1079.1
散金: 30
帖子: 105
在线: 70.6小时
虫号: 896951
注册: 2009-11-08
性别: GG
专业: 工业工程与管理

[求助] 电子商务个性化推荐方向投稿交流

各位大神，本文现在刚刚写完了一篇英文文章，是关于电子商务网站用户个性化推荐的文章。想投稿一篇sci期刊，不知道诸位有没有审稿周期快，录用可能性高的相关期刊推荐。个人感觉投稿 “IEEE Transactions on Knowledge And Data Engineering”期刊难度稍大，不知道有没有合适的期刊推荐以下。拜托诸位了！！！
文章名称
Predicting E-Commerce Users’ Visiting Paths Based on All-kth Markov Model Considering Browse Time
摘要
Accurate prediction of users’ visiting paths is critical for e-commerce companies to improve the performance of their recommendation systems. Prior research has adopted many features of users’ past behaviors in building models for predicting their future visiting paths, but has overlooked one important parameter, namely, users’ browse time. This paper proposes a prediction model that incorporates users’ browse time for the sake of improved prediction accuracy. Users’ browse time is split into several segments by adopting the top-v nearest splitting time segments to analyze users’ next visiting path. We build an all-kth Markov model incorporating browse time to predict users’ visiting paths. Prediction accuracy of this model improves as order (i.e., k) increases and reaches the peak value when k is from 5 to 7. Incorporating browse time into the all-kth Markov improves its performance in predicting users’ visiting behaviors, by decreasing the mean absolute error of prediction from 3.45 to 3.19. Further, we analyze the relationships between the number of splitting time segments and the number of adopting time segments. A case based on 300,000 users of a famous e-commerce website in China suggests that there is a power relationship between the number of the optimal adopting time segments and the number of splitting time segment. These findings have important implications for customer behavior analysis theory and practice.

回复此楼