Venture capital project estimation based on cost sensitive support vector machines

摘要

The success of venture investment lies in making reasonable investment decisions to the candidate project, and what is critical to this investment decision-making is to conduct an appropriate project assessment. In this paper, we propose an interval learning based cost-sensitive support vector machines (CISVM) approach to improve the classification model of a venture investment project, it can overcome the disadvantage of the conventional cost-sensitive algorithms that cannot work with the cost interval in the project. The experimental results of the learning cost interval over a real venture investment data set verified the effectiveness of the proposed method.