| At present,with the aging of China’s population,China’s senior care industry is facing increasingly severe trends and the operational problems of senior care institutions are coming to the fore.Meanwhile,with the development of data mining techniques such as clustering and increasingly common tools for data analysis support,coupled with data operation solutions integrating multiple technologies,the construction of a comprehensive elderly care system covering nursing,medical and recreational care has been promoted.This paper proposes a clustering approach based on electricity data for the abnormal operation of senior living institutions for identifying and solving challenges that may need to be overcome in the operation and management of senior living institutions.Using senior living institutions as the object of analysis and research,we use electricity data such as electricity consumption,electricity bills,electricity consumption categories,and industry categories from the electricity marketing category as the basis,and converge and fuse data related to senior living institutions provided by civil affairs departments,from overall electricity consumption,regional The analysis model is mainly based on STL time series decomposition method to obtain longterm trend data of elderly institutions to assist civil affairs departments in decision making,and K-Means clustering algorithm and isolated forest algorithm are used to rank suspected abnormal elderly institutions to obtain a list of abnormal institutions.The list of institutions with abnormal operation can be screened out to assist the civil affairs department to make relevant operational decisions and analysis.Electricity data-based operation analysis of senior care institutions is a data product to help senior care institutions refine their management,providing decision support to improve the operation capability of senior care institutions,enhance the level of civil administration and promote the high-quality development of people’s livelihood.Through machine learning methods to mine the potential laws in electricity marketing data and build an abnormal identification model for senior care institutions to identify abnormal electricity consumption behavior of senior care institutions,it improves the efficiency and accuracy of monitoring,effectively saves comprehensive costs such as manpower,and optimizes the resource allocation of various institutions.It can provide residents with senior care service comparison and assist civil affairs departments to optimize the operation and management of senior care institutions.The advantages and shortcomings of the research method are also summarized at the end of the article based on the existing data,optimized model and analysis results,and the model parameters can be adjusted to optimize model training by optimizing algorithms such as outlier screening. |
