Research On Recommendation System For Equipment Fault Maintenance Experts Based On Text Analysis

With the vigorous development of the industry,many large-scale modern equipment has emerged.These large-scale equipment can improve industrial production efficiency and promote social development.It can be said that human development is increasingly inseparable from the support of large-scale advanced equipment.However,the more powerful the equipment,the more complex and expensive the machine structure,the longer the use of large modern equipment,the greater the number of faults exposed,the more extensive,complex and changeable.Once the equipment has a problem that cannot be solved by itself,the only way is to find a professional maintenance expert to repair the equipment.It is a problem to recommend which maintenance expert to repair the equipment faults accurately and efficiently.The problem is used as the research object,using the relevant data generated in the historical maintenance tasks to carry out research from the field recommendation of fault maintenance experts and the personalized recommendation of maintenance experts,etc.The purpose is to find the optimal fault maintenance expert recommendation according to the needs of users Program.In order to solve the recommendation problem in the field of fault maintenance experts,this paper proposes an improved BERT fault text classification model,which can be used to classify fault text description information submitted by users,and then find all good repair experts in the field of failure.The domain recommendation problem is transformed into a text classification problem,the recommendation task of the recall layer is completed,and a candidate set of maintenance experts is generated to prepare for the personalized recommendation of maintenance experts.In order to solve the problem of personalized recommendation of fault maintenance experts,this paper proposes a personalized recommendation model for fault maintenance experts based on XGBoost that integrates users,maintenance experts,and their interaction characteristics.The score of each candidate fault maintenance expert generated by the layer,the score of each maintenance expert is sorted according to Top-k,and the first few fault maintenance experts with higher scores are selected as the final recommendation result.The recommendation method can effectively ensure the personalized recommendation for different user needs,complete the recommendation task of the ranking layer,and effectively improve the accuracy of the recommendation results.Based on the above research results,an expert recommendation system for equipment fault maintenance is designed and implemented.The system provides users with the intelligent management function of fault report work orders and the field and personalized recommendation function of maintenance experts.Users can use the web to interact with the system,which improves the user’s data management efficiency and helps users make reasonable decisions.