Research And Application On Screening Optimization Of Two-Piao Assessment System Based On Content Recommendation

The development of China's national economy is inseparable from the support of sufficient power,and thermal power generation is still the main way of generating electricity in China.In recent years,the research of intelligent recommendation algorithms has been deepened,and the problem of obtaining effective information under massive data has been effectively solved.In this paper,the problems of “TwoPiao” training for employees in thermal power generation enterprises,combined with the intelligent recommendation algorithm,consider the actual project realization of the "Two-Piao" expert assessment system,and propose a screening optimization algorithm based on content-based ticket exercises.In this paper,the commonly used recommendation algorithms are studied.By analyzing the characteristics of the operation tickets used by thermal power generation enterprises,the “Content-based Recommendation Algorithm” is selected as the basis.The Chinese word segmentation device is used to perform Chinese word segmentation on the operation items and operation tasks,and the description phrases of all operation tickets are obtained.The keywords of each operation ticket are then filtered by the TFIDF algorithm.Comparing and analyzing the advantages and disadvantages of One-Hot coding vector and Word2 Vec tool generation vector,we choose to use the operation ticket text as the corpus to train the Word2 Vec model,and initially obtain the space vector of the operation ticket object.The different weights of the keyword vectors of different parts of speech are given,and the space vector corresponding to the corresponding ticket is obtained.Considering the corresponding engineering implementation and system upgrade,the obtained operation ticket space vector is clustered by K-means++,and further clustered into 3 categories based on the ticket category.Finally,the user interest model generation strategy commonly used in the recommendation algorithm is analyzed.Considering the actual use of the operation ticket of the employees of the thermal power generation enterprise,the user's current interest ticket is selected as the benchmark,the user interest model is established,and the complete ticket exercise screening recommendation is completed.The design of.The actual test proves that for the reference operation ticket,the algorithm can effectively filter out the operation ticket with high similarity to the user,which is convenient for the user to strengthen the learning of relevant knowledge points.