Application of improved k-modes clustering algorithm in collaborative filtering employment recommendation algorithm
In order to provide accurate personalized employment recommendation service for college graduates,the improved k-modes clustering algorithm based on dynamic weight interdependence distance is applied to collabora-tive filtering recommendation algorithm.It is defined that the distance between attributes of different sample points is equal to the weighted sum of the internal distance of the attribute values and the external distance between the at-tributes.When the initial cluster centroid is selected,the distance between the sample point and the cluster centroid and the combined weight of the cluster density are dynamically adjusted,and the radius of the cluster density calculation formula is dynamically set,and the initial cluster centroid is selected according to the prob-ability value of the sample points.The student cluster and position cluster meeting the accuracy are obtained by iterative calculation and optimization.The student-post matrix is constructed to calculate the similarity between fresh graduates and previous graduates,and the similarity between former graduates and entry posts.The combina-tion of fresh graduates and job clusters whose similarity exceeds the threshold is selected as matching pair for matching.The matching information is arranged in descending order to form a matching list,and bidirectional recommendation and information push are carried out according to the matching list.To provide information guid-ance and technical support for university employment recommendation and guidance.
NETL
NSTL
万方数据
