Resource association method based on crowd intelligence aggregation convergence

Abstract

The invention discloses a resource association method based on crowd intelligence aggregation convergence. Firstly, public participation-oriented learning resource labeling tasks are formulated, and participants are mobilized to mark relevance degrees of knowledge, then deception-type participants are screened out through evaluating task completion situations of the participants, and are marked asunqualified users; remaining higher-quality label results are collected after the unqualified users are filtered out, mark classifications made by participants are utilized and an EM algorithm is used to continuously iterate to calculate each participant error rate pi and each relevance T, iteration is continually carried out through the EM algorithm until convergence is reached, and thus the relevance degrees between knowledge points are estimated; and a label result set of the participants is obtained by collection, association degrees between resources are calculated, and finally, association data are stored to a server. Relevance between the knowledge resources is obtained, the deception-type participants can be effectively found and replaced in time, crowd intelligence label qualityis improved, and technical support is provided for deeply mining association between the learning resources.

Description

technical field [0001] The invention belongs to the field of educational technology, and relates to a method for association evaluation of learning resources, in particular to a method for association of learning resources based on convergence of crowd intelligence. Background technique [0002] In the era of digital learning (e-learning), the learning trajectory of learners no longer presents a linear learning method as in the past. At the same time, the number of educational resources continues to increase, but there is a lack of good organizational relationships among resources. A lot of inconvenience. [0003] At present, automatic machine processing algorithms are used to help learners calculate resource associations. However, for heterogeneous synonyms with the same semantics and inconsistent writing, such as implicit knowledge associations of learning resources, the accuracy of traditional machine processing methods is often limited. , it is impossible to dig deep in...

AI Technical Summary

Problems solved by technology

[0002] In the era of digital learning (e-learning), the learning trajectory of learners no longer presents a linear learning method as in the past. At the same time, the number of educational resources continues to increase, but there is a lack of good organizational relationships among resources. a lot of inconvenience

[0003] At present, automatic machine processing algorithms are used to help learners calculate resource associations. However, for heterogeneous synonyms with the same semantics and inconsistent writing, such as implicit knowledge associations of learning resources, the accuracy of traditional machine processing methods is often limited. , it is impossible to dig deep into the relationship between knowledge resources; relying on the manual annotation of a few experts has huge manpower and time costs, and a more effective way is needed to solve this problem

Images