Alibaba Group Holding Limited

Congzhi Wang

Hao Wang

Jingjun Ye

Xiaolong Ma

The present invention is a method and a system for determining junk information related to the fields of information filtering, and can solve the problem of improperly predefined junk information. The method includes: pre-defining junk information keywords in data samples; computing keyword scores for each of the keywords; simulating filtering of the data samples using the keyword scores to obtain a filtering index value; determining whether the filtering index value satisfies an evaluation index; if not satisfied, revising the keywords or the keyword scores and repeating the simulating to obtain a new filtering index value; and if satisfied, terminating the revising. The disclosed method can help a user more reasonably define junk information keywords and keyword scores, and determine whether information is junk information based on the keywords and corresponding scores, thereby improving efficiency of information filtering. The invention can be widely adopted in a variety of applications and systems such as feedback systems, guestbook systems, forums and junk mail processing.

Growing community of inventors

Zhejiang, China

Congzhi Wang