A behavior prediction method and device

Abstract

The invention provides a behavior prediction method and device. The method comprises the steps of: acquiring historical data sets of a plurality of users, wherein each historical data set comprises a plurality of pieces of historical data, and the historical data comprises a target behavior; processing all the historical data to obtain maximum frequent item sets corresponding to all the historical data, wherein each maximum frequent item set comprises a plurality of elements, all the elements are sorted according to a support degree sequence from small to great, each element comprises a plurality of pieces of historical data, and all the historical data in each element is sorted according to a support degree sequence from small to great; according to the target behavior and each piece of historical data in the maximum frequent item sets, determining a plurality of effective association rules for deducing the target behavior; and according to a historical data set of a target user and each effective association rule, determining a probability of executing the target behavior by the target user. The behavior prediction method and device utilize an improved association rule algorithm to carry out predication, and are high in prediction accuracy, simple to operate and excellent in practicality.

Description

technical field [0001] The present invention relates to the technical field of behavior prediction, in particular to a behavior prediction method and device. Background technique [0002] Currently, our society is undergoing tremendous changes in terms of economy, politics, science and technology. Although the development of society has brought a lot of convenience to our life, all kinds of illegal and criminal behaviors often appear around us. These high-tech, gang-oriented, and intelligent criminal behaviors have seriously affected our normal life order. In the long run, it has become a major obstacle to the harmonious development of society. Over the years, a large amount of criminal data has been accumulated in the prison reform information management system. However, how to make good use of these existing data on criminals to achieve effective analysis and control of criminal behavior is a major problem facing prison managers, and it is also a subject that the prison...

AI Technical Summary

Problems solved by technology

[0004] The inventor found in the research that the recidivism prediction method based on the decision tree algorithm in the prior art is limited by the multi-valued bias problem of the decision tree algorithm itself, resulting in relatively single prediction results and low prediction accuracy, while The re-forecasting method based on comprehensive science has a complicated operation process and poor practicability

In addition, there is no perfect solution for the re-prediction of criminal behavior that satisfies both high prediction accuracy and good practicability.

Images