Company classification using machine learning
作者：
Highlights：
• Combining t-SNE and spectral clustering achieves compelling company classification.
• The decision engine for optimal grouping allows incorporation in various frameworks.
• Selecting the group size with cross-validation improves the resulting classification.
• Applied to portfolio optimization, our approach results in a Sharpe ratio of 1.5 p.a.
• Applied to company valuation, our approach achieves high valuation accuracy.
摘要
•Combining t-SNE and spectral clustering achieves compelling company classification.•The decision engine for optimal grouping allows incorporation in various frameworks.•Selecting the group size with cross-validation improves the resulting classification.•Applied to portfolio optimization, our approach results in a Sharpe ratio of 1.5 p.a.•Applied to company valuation, our approach achieves high valuation accuracy.
论文关键词：Classification,Unsupervised learning,t-SNE,Spectral clustering,Portfolio optimization
论文评审过程：Received 20 May 2020, Revised 5 January 2022, Accepted 20 January 2022, Available online 5 February 2022, Version of Record 10 February 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.116598
