A Generation Method Of Mobile Users’ Tags Based On Time Features

As mobile Internet advances, the society today has been entering the era of exploding information and data. To find valuable information in big data, such as people’s individual features, and make better plans for individual recommendations, has become one of the most important purposes of most companies. In nowadays, most individual features of users are described by feature tags, which are made up of a few of words. However, with the quick development of data from users and the scale of users, the traditional method to describe people’s individual features can not satisfy the purpose of most companies. A more detailed method to describe people’s individual features is required urgently.This paper presents a new describing method for people’s individual features which is based on the traditional one and added with the time features. The time features of people’s individual features are made of short-long feature, time-period feature and day-cycle feature. The short-long feature can present the track of behavior of one user. A larger long-feature of a user tag means this user is interested on it during a long period of time. A larger short-feature of a user tag means this user is interested on it only during a short period of time.In the experiment, we compare the recommending results of user tags with time features and those without time features, and find out that the coverage rate of recommendation of the results of user tags with time features is 77% which is better than the other one which is only 68%. So, we conclude that the description of user tags’time features is reasonable and meanful.