User behavior information analysis method and device thereof

Abstract

An embodiment of the invention discloses a user behavior information analysis method and a device thereof. The user behavior information analysis method comprises the steps of: according to user logging-on information, extracting attribute values of preset attributes, wherein the preset attributes comprise an user attribute and commodity kind to which a commodity that is purchased by the user last time belongs; determining a combined attribute KEY according to the attracted attribute value of each preset attribute, inquiring in a database for obtaining a value which corresponds with the combined attribute KEY, wherein the value which corresponds with the combined attribute KEY comprises number of historical purchasing times of each commodity kind; and according to the number of historical purchasing times of each commodity kind in the value which corresponds with the combined attribute KEY, determining the first N commodity kinds with maximal number of historical purchasing times as latent consumption objects of the user. The user behavior information analysis method according to the embodiment of the invention can improve accuracy in determining the latent consumption behavior of the user.

Description

technical field [0001] The present application relates to the field of Internet technology, in particular to a method and device for analyzing user behavior information. Background technique [0002] With the development of the Internet, it is often necessary to analyze user behavior information in order to predict the user's future behavior, and then determine Internet information distribution or processing strategies based on the prediction results. [0003] At present, user behavior analysis is usually done by building a model. Specifically, first collect samples from historical data, build a prediction model based on the collected samples, and then use the prediction coefficients in the prediction model, such as probability coefficients, support coefficients, or Relationship weight coefficients, etc., to predict the potential behavior of users. [0004] For example, in the field of e-commerce, when a prediction model is established based on the association rules between...

AI Technical Summary

Problems solved by technology

[0005] It can be seen that when analyzing user behavior by establishing a model, since the model is established based on certain rules, such as establishing a model based on association rules between products, the rules on which the model is based may not fully cover the consumption behavior of all users , so the established model has a certain one-sidedness and randomness, it cannot be guaranteed to be applicable to all users, and the accuracy rate of determining the potential consumption behavior of users is low

[0006] Moreover, considering the time dimension, once the model is established, unless the sample is re-selected and the model is re-established, the prediction coefficient in the model is fixed, so it cannot adapt to the latest consumption behaviors such as changing consumption habits and consumption trends. Information, resulting in a decline in the accuracy of determining the potential consumption behavior of users

Images