User recommendation method, device, equipment and medium based on user time

Abstract

The present invention relates to the technical field of group portraits. The present invention discloses a user recommendation method, device, equipment and medium based on user duration. The method includes: obtaining log data to be counted and performing log interpretation on the log data to be counted to obtain interpretation data; performing instance deduplication on the interpretation data to obtain deduplication data; using Mapreduce to perform parallel minute aggregation processing on the deduplication data through a distributed balancing strategy to obtain multiple minute-level duration data; performing granular dimension-upgrading aggregation processing on all minute-level duration data to obtain daily duration data; performing user time profiling on the daily duration data to obtain user recommendation results. Therefore, the present invention realizes automatic analysis of daily duration data based on log data, automatically profiling the user type with longer duration, and obtaining user recommendation results based on the user type, thereby improving the accuracy and precision of the duration, making corresponding measures for targeted users in the future, and effectively improving the user return rate.

Images