Search method, search apparatus and non-temporary computer-readable storage medium for text search
AssigneeBeijing Sankuai Online Technology Co., Ltd
InventorsMing Liu, Dayao Chen, Mengmeng Pang, Tao Feng, Zhizhao Zeng, Yongchao Wei, Wenbin Pan
Life of this patent
Estimating the term…
Abstract
A search method includes: determining at least one first search policy matching a to-be-searched query text, where each first search policy corresponds to at least one first text index field and a search weight matching the first text index field; performing a search operation for the query text based on each first text index field corresponding to each first search policy and a search weight matching the first text index field; and merging and outputting search results of all the search operations.
Claims12
A search method, comprising:
- obtaining at least one first search policy matching a to-be-searched query text, wherein each of the at least one first search policy corresponds to at least one first text index field and a search weight matching the at least one first text index field;
- determining the at least one first text index field corresponding to the at least one first search policy and the search weight matching the first text index field, wherein the determining comprises:
- obtaining a search log corresponding to the at least one first search policy;
- performing an iterative calculation on an average weight of each second text index field corresponding to the at least one first search policy according to a hit score of a query text in the search log corresponding to the at least one first search policy in the second text index field in a search material;
- calculating a normalized weight value of each second text index field corresponding to the at least one first search policy based on the average weight of each second text index field corresponding to the at least one first search policy;
- assigning a respective second text index field with a respective normalized weight value greater than a preset threshold as the at least one first text index field corresponding to the at least one first search policy;
- and
- assigning a respective normalized weight value of the respective second text index field as the search weight matching the at least one first text index field;
- performing a search operation for the to-be-searched query text based on the retrieved at least one search policy, the at least one first text index field, and the determined search weight;
- performing a search operation for the to-be-searched query text based on a second search policy corresponding to all text index fields of the search material, wherein search weights of all the text index fields are identical;
- and
- merging and outputting search results of all the search operations.
A search apparatus, comprising:
- a processor;
- and
- a non-temporary computer-readable storage medium, wherein
- the non-temporary computer-readable storage medium stores a machine executable instruction executable by the processor, and the machine executable instruction causes the processor to perform the following operations:
- obtaining at least one first search policy matching a to-be-searched query text, wherein each of the at least one first search policy corresponds to at least one first text index field and a search weight matching the at least one first text index field;
- determining the at least one first text index field corresponding to the at least one first search policy and the search weight matching the first text index field, wherein the determining comprises:
- obtaining a search log corresponding to the at least one first search policy;
- performing an iterative calculation on an average weight of each second text index field corresponding to the at least one first search policy according to a hit score of a query text in the search log corresponding to the at least one first search policy in the second text index field in a search material;
- calculating a normalized weight value of each second text index field corresponding to the at least one first search policy based on the average weight of each second text index field corresponding to the at least one first search policy;
- assigning a respective second text index field with a respective normalized weight value greater than a preset threshold as the at least one first text index field corresponding to the at least one first search policy;
- and
- assigning a respective normalized weight value of the respective second text index field as the search weight matching the at least one first text index field;
- performing a search operation for the to-be-searched query text based on the retrieved at least one search policy, the at least one first text index field, and the determined search weight;
- performing a search operation for the to-be-searched query text based on a second search policy corresponding to all text index fields of the search material, wherein search weights of all the text index fields are identical;
- and
- merging and outputting search results of all the search operations.
A non-temporary computer-readable storage medium, storing a machine executable instruction, wherein when invoked and executed by a processor, causes the processor to perform actions of:
- obtaining at least one first search policy matching a to-be-searched query text, wherein each of the at least one first search policy corresponds to at least one first text index field and a search weight matching the at least one first text index field;
- determining the at least one first text index field corresponding to the at least one first search policy and the search weight matching the first text index field, wherein the determining comprises:
- obtaining a search log corresponding to the at least one first search policy;
- performing an iterative calculation on an average weight of each second text index field corresponding to the at least one first search policy according to a hit score of a query text in the search log corresponding to the at least one first search policy in the second text index field in a search material;
- calculating a normalized weight value of each second text index field corresponding to the at least one first search policy based on the average weight of each second text index field corresponding to the at least one first search policy;
- assigning a respective second text index field with a respective normalized weight value greater than a preset threshold as the at least one first text index field corresponding to the at least one first search policy;
- and
- assigning a respective normalized weight value of the respective second text index field as the search weight matching the at least one first text index field;
- performing a search operation for the to-be-searched query text based on the retrieved at least one search policy, the at least one first text index field, and the determined search weight;
- performing a search operation for the to-be-searched query text based on a second search policy corresponding to all text index fields of the search material, wherein search weights of all the text index fields are identical;
- and
- merging and outputting search results of all the search operations.
Dependencies are inferred from the claim text.