Recommendation method and system for local low-rank matrix approximation based on implicit feedback information

Abstract

The invention discloses a local low-rank matrix approximation recommendation method and system based on implicit feedback information, and belongs to the technical field of recommendation. The invention starts from the implicit feedback information ignored in the user information, uses the local low-rank matrix approximation principle, uses the implicit feedback information to optimize the feature extraction model, and extracts the single-dimensional multi-dimensional preference vector between the user and the item; Preference information, giving a more accurate user rating recommendation method. The recommendation system of the present invention includes a data engine module, a multi-dimensional feature extraction module, a score prediction module, a collaborative recommendation module and an information update module; through the collaborative work of each module, the multi-dimensional recessive factors of users in different environments are extracted, and the overall The accuracy and personalization of the system recommendation results have high generality and are suitable for most existing recommendation scenarios.

Description

technical field [0001] The invention belongs to the field of recommendation technology, and more specifically, relates to a recommendation technology based on local low-rank matrix approximation based on implicit feedback information. Background technique [0002] The development of Internet technology has brought us many conveniences, but it has also brought us many difficulties. The most famous of these is the “information overload” problem. Due to the convenience of the Internet, all our information is interacted on the Internet, and all data are stored in cloud databases, resulting in a geometric multiple of network data, which makes it difficult for people to find the information they need on the Internet. [0003] The search engine is an important invention, which allows people to retrieve matching information based on keywords from massive data, and alleviates the pressure on users caused by excessive data volume to a certain extent. However, it ignores that differe...

AI Technical Summary

Problems solved by technology

If these hidden factors are ignored only based on the analysis of the user's historical data, it is likely that the change speed of the recommendation results will be slower than the change speed of the user's preferences, resulting in a lag in the recommendation results; Preferences, as ambient noise, processed, ignored

Images