888贵宾会手机版

888贵宾会手机版:学术报告

发布时间:2015-04-14 文章作者:浏览次数:12

    报告人:闵帆(报告人简介:闵帆,博士、漳州师范学院教授、博士生导师。主要从事关于粒计算与代价敏感学习问题的研究,近年来发表论文100余篇,其中3篇入选ESI高被引用论文,单篇最高引用136次。带领西南石油大学机器学习团队http://scs.swpu.edu.cn/smale从事代价敏感学习、推荐系统等研究。)      

时间:2015年4月14日下午2点至6点

       地点:888贵宾会手机版会议室

    讲座题目:Cost-sensitive learning through granular computing Abstract: Cost-sensitive learning is among the top-10 challenging problems in machine learning and data mining. It refers to the learning process with various types of costs, including test cost, misclassification cost, teacher cost, and so on. For classification, we try to minimize the misclassification cost (also called risk) instead of the error. For feature selection, we try to minimize the test cost instead of the number of features. For the tradeoff between the cost and the benefit, we try to minimize the sum of test cost and misclassification cost. For active learning, we try to consult the expert with teacher cost for appropriate data. In this talk, we will discuss granular computing approaches to cost-sensitive learning.

       

    
888贵宾会手机版-贵宾会app手机版