| As an aging population that continues to grow,the elderly care service industry has been prospering.In contrast to the excessive existing options,the services that each elderly person and their family are really interested in are very limited,which causes customers have to spend a lot of time and effort to find the suitable service.In addition,due to the improving living standards and individual differences,the demands tend to be more diverse and personalized.Therefore,actively recommending elderly care services that satisfy the individual demands and interests of elderly people is becoming an important research problem in modern elderly care service industry.Applying recommendation methods in elderly care services can solve the above issues.Effective recommendations can not only improve customer satisfaction but also promote high-quality services.Therefore,in addition to integrating elderly care service resources,this paper studies the recommendation methods in the field of elderly care services.Firstly,regarding the lack of unified description of services,research is carried out from three aspects: classification,content and quality of service.Classification theories are combined with real customer demands to better classify elderly care services;Service contents are vectorized by extracting key words using text processing algorithms;Quality evaluation systems are established through analytic hierarchy process to ensure the quality of elderly care service.A personalized tree model is also build and the elderly care service problem is formalized.Secondly,to study elderly people’s individual requirements and preference,a multi-dimensional elderly-oriented user profile database is established.We obtain and model customers’ static and dynamic features,extract their demands on elderly care services,and design labels,weights and time decay functions to dynamically describe the demands and interests of elderly people.Thirdly,users’ demands and preference presented in their user profile are combined with the description method of elderly care services to study the recommendation algorithms.Content-based recommendation algorithm and matrix factorization based on user profile are proposed to address different cases.These two algorithms are both implemented to formalize and approach the elderly care services recommendation problem,the performances are compared and analyzed through experiments.Finally,on the basis of recommendation methods,an personalized elderly care service oriented recommendation system is developed in terms of system architecture,functionality,database and technology.The key functions are implemented and tested. |
