A social search evaluation method based on friend clustering in lbsn

Abstract

The invention discloses a friend clustering-based social search evaluation method for an LBSN. Multi-dimensional characteristics such as location-based information, contact person information and the like are extracted from a Foursquare real data set; a friend clustering-based KNN search algorithm is proposed; a reverse index-based search engine is designed; and in combination with factors such as a distance and the like, a search result is more accurate and the search speed is increased. For enabling the search result to be more accurate, firstly, friends are clustered on the basis of researching user friends. The LBSN belongs to a heterogeneous network, and the data set is relatively sparse, so that data can be denser by clustering; singular points are eliminated, so that adverse influence caused by data sparsity is reduced; secondly, in design of the search algorithm, on the basis of considering conventional social contact influence, two indexes including professional relevance and distance are added, namely, a comprehensive search score, a social contact score and a distance score are considered; and finally, the three indexes are integrated, a linear planning model is built and trained, and the search result is obtained, so that a user is satisfied with the search result.

Description

technical field [0001] The invention relates to an evaluation method, in particular to a social search evaluation method based on friend clustering in LBSN, and belongs to the technical field of social search. Background technique [0002] The development of Online Social Networks (OSNs) has brought great convenience to people's daily life. Today, billions of users are active on OSNs every day, generating a large amount of social information. Gradually, people prefer to search for information through OSNs instead of traditional search engines, and social search emerges as the times require. Due to the shortcomings of traditional search methods such as low precision rate, long user screening time, and consistent search results, in the context of user personalized search, social search relies on traditional search principles and combines user social information to generate personalized search results. search results to improve search accuracy. In particular, the emergence of...

AI Technical Summary

Problems solved by technology

Some studies pay more attention to the improvement of search speed, and improve the search speed by improving the index structure, ignoring the sparseness of the data set, which affects the search accuracy; some studies analyze specific events that occur at the location point, such as earthquakes, fires, etc., Realize the search for events that occur at specific locations, but lack the generalization of the search object field, the field is relatively single, and cannot better meet the actual needs; some studies are due to the subjective evaluation methods such as similarity metrics and the overload of social search systems Large, resulting in room for improvement in search performance

Images