Inquiry-related multi-ranking-model integration algorithm

Abstract

The invention discloses a brand new inquiry-related multi-ranking-model integration method. The brand new inquiry-related multi-ranking-model integration method comprises the following steps of: establishing sub-ranking models for each inquiry and inquiry-related document; performing vectoring representation on the sub-ranking models so as to convert a plurality of inquiry-related ranking models into characteristic data and integrate the plurality of ranking models; establishing a new loss function as an optimization target at an inquiry level and a sample level by using a ranking support vector machine as the sub-ranking model; adjusting weight among losses generated by different inquiries by using the loss function; and providing a multiple inquiry-related ranking support vector machine fusion algorithm. Compared with the traditional model, the inquiry-related ranking models can achieve better properties when the inquiry-related multi-ranking-model integration algorithm provided by the invention is applied to actual tasks. The multi-model fusion algorithm provided by the invention can be applied to ranking learning, can also be applied to multi-element classification, sequence labeling and the like, and has a wide application prospect in the fields of information retrieval, network search and the like.

Description

technical field [0001] The invention relates to the fields of information retrieval and machine learning, in particular to a plurality of sorting support vector machine model fusion algorithms related to queries. Background technique [0002] Learning to rank is a hot topic in the field of information retrieval and machine learning. Information retrieval refers to finding a subset of information related to a given query from a large number of document collections, and is an important means of processing massive text information. In the vast majority of current information retrieval systems, the retrieved information (such as documents, etc.) is returned to the user in a sorted manner. Therefore, the core problem of information retrieval model research is how to efficiently sort information . The purpose of ranking learning is to find a ranking that can accurately predict unknown data The decision function for label y [0003] Traditional learning-to-rank methods can b...

AI Technical Summary

Problems solved by technology

[0008] However, neither the list-based learning-to-rank method nor the learning-to-rank method that directly optimizes the evaluation metric takes the inter-query differences into account during the modeling process.

Although the query-related ranking learning algorithm considers the query characteristics more than the traditional single model, there are still two problems: one is that the selection process of the characteristics is separated from the establishment of the ranking model, which is not reflected in the ranking model loss; It is to train multiple models, which is not convenient for the processing of the final sorting results

Images