Entity expanding method of network list

Abstract

The invention provides an entity expanding method of a network list. The method comprises the steps of selecting seed lists to seed groups by calculating a semantic matching value between the networklist and a query list; selecting the group list with the maximum list potential to add into the seed groups, wherein the group list and the seed groups accord with consistency matching relations and are used for rising the coverage rate of the seed groups, and when the coverage rate reaches a set threshold value, the seed groups are considered as consistency groups meeting the given coverage rate;when nodes of the consistency groups are regarded as an answer list needed by entity expansion, using the answer list to establish a final event list expanding the entity consistency. Searching for answers, the entity expanding method of the network list brings in a concept of a consistency matching relation, the consistency of the answer list is improved, and therefore the expansion of multiplesearching lists can be adapted, so that the result consistency is guaranteed, and at the same time, the high-precision and reliability of the result are also ensured.

Description

technical field [0001] The invention relates to the technical field of structured data integration on webpages, in particular to a network table entity extension method. Background technique [0002] Users usually want to obtain the information they are interested in, and they can use a large number of tables on the network as information sources, and realize it through entity expansion. Existing techniques assume that web forms are entity-attribute binary relationships. For tables with multiple columns of attributes to be expanded, the existing technology first splits these tables into several entity-attribute binary relations, and then aggregates the results of separate expansion into a complete answer. The result of this is that the semantics of the table are split during the splitting process, and the resulting table composed of split binary relations inevitably has the disadvantages of entity inconsistency and low accuracy. [0003] The InfoGather system proposed by M...

AI Technical Summary

Problems solved by technology

The premise of adopting this strategy is that the attributes in the network table are not related to each other. This assumption obviously ignores the connection between the attribute columns, resulting in the splitting of the table semantics, resulting in low accuracy of entity expansion and inconsistent entities. question

[0004] Most of the network tables are n-element tables. Using existing technology to split them will destroy the semantics of the tables, resulting in inconsistencies between spliced ​​entities and attributes

The web form is not standardized, there are problems such as missing column labels, and it is impossible to judge the matching relationship between tables based on the column labels

Entities are ambiguous. Entities with the same name may have different semantics. Only relying on entities to judge the matching relationship between tables will lead to semantic conflicts between matching tables.

Images