Method and method for recommending data

Abstract

The invention relates to a method and system for recommending data, wherein the method comprises the steps of obtaining a conditional probability value of a user to each data category according to historical records of the user to data categories, obtaining a read trend value of all users to each data category according to the historical records of all the users to the data categories, obtaining a historical preference value of one user to each data category in a time period according to the historical records of the user to the data categories, obtaining a predicted preference value of the user to each data category in a time period according to the conditional probability value, the read trend value and the historical preference value, and ranking the predicted preference values in the same time period from high to low, obtaining first M data categories highest in the predicted preference value and recommending the obtained M data categories to the user. The method and the system are capable of determining the data categories in which the user are interested, and obtaining data from the data categories and recommending the data to the user, and as a result, the method and the system are capable of recommending interested contents for the user more effectively, reducing invalid browsing of the user and improving the read satisfaction of the user.

Description

technical field [0001] The invention relates to a data recommendation method and system. Background technique [0002] With the popularity of 3G networks and the rise of smart phones and tablet computers, users can access information anytime, anywhere. Surveys show that users usually have a strong need to browse, read information, and pass free time through smartphones and tablet computers. [0003] Existing data recommendation methods or systems usually have the following problems: [0004] 1. When users read data such as information, their purpose is not to discover information and acquire knowledge, but more often they will have such an attitude: "Give me something interesting." Therefore, if you pass every The information itself models the user's interest model, and then recommends information to the user according to the interest model, but the recommended information is often limited to the content that the user already knows; [0005] 2. Users' points of interest a...

AI Technical Summary

Problems solved by technology

[0004] 1. When users read data such as information, their purpose is not to discover information and acquire knowledge, but more often they will have such an attitude: "Give me something interesting." Therefore, if you pass every The information itself models the user's interest model, and then recommends information to the user according to the interest model, but the recommended information is often limited to the content that the user already knows;

[0005] 2. Users' points of interest are easily affected by surrounding emergencies. For example, many users are called "big game fans", that is, these users have never paid attention to football before, but due to the recent "European Cup" in full swing and media publicity Discuss with colleagues and friends around, so that these users will also be willing to accept football information. The existing recommendation methods or recommendation systems usually cannot grasp the changes in user interests in this case and make corresponding recommendation adjustments;

[0006] 3. The user's points of interest may change in different time periods. For example, the user may be more willing to receive news and other information on the way to work every day, and the user is more inclined to receive relaxed and pleasant information that does not require thinking. Eliminate the fatigue of the day, and in the dead of night, the user may want to read some in-depth reports, etc. The existing recommendation methods or recommendation systems usually cannot take into account the user's interest changes in different time periods

Images