Clustering stock price time series data to generate stock trading recommendations: An empirical study
作者：
Highlights：
• Trading recommender system based on mining historical stock price data is proposed.
• Regression tree and Self Organizing Maps are used to generate temporal clusters.
• Temporal clusters are then used to generate trading recommendations.
• 16 recommender system variants are evaluated on US, UK, India and Brazil stocks.
• Proposed recommenders are capable of generating profitable trade recommendations.
摘要
•Trading recommender system based on mining historical stock price data is proposed.•Regression tree and Self Organizing Maps are used to generate temporal clusters.•Temporal clusters are then used to generate trading recommendations.•16 recommender system variants are evaluated on US, UK, India and Brazil stocks.•Proposed recommenders are capable of generating profitable trade recommendations.
论文关键词：Stock,Trading,Recommender,Clustering,Time-series
论文评审过程：Received 30 May 2015, Revised 17 August 2015, Accepted 1 November 2016, Available online 2 November 2016, Version of Record 21 November 2016.
论文官网地址：https://doi.org/10.1016/j.eswa.2016.11.002
