| dc.description.abstract | Aging is a challenge facing the global society, which leads to labor shortage, social cost increase, and lack of medical energy and resources. The proper care of the elderly is the primary force behind the stability of a society. However, the proper care of the elderly has always been an issue that needs to be addressed in various fields. The proper care of the elderly is not only about food and clothing, but also about the spiritual foundation that requires assistance and efforts from all parties. Therefore, this study uses data-driven technology to construct an interpersonal community network for the elderly based on the local friendship recommendation model, to increase the social bonding of the elderly, to enhance their social participation, to expand their friendship circle, to promote the activity of the aging process, and to reduce the possibility of lonely aging. In the first stage, we used deep learning techniques to analyze the personality traits implied by the elderly′s life data, in order to overcome the time consuming difficulties in assessing personality traits by traditional questionnaires. In the second stage, we use machine learning technology to combine the personality traits, interests and hobbies of the elderly with the geographic information of routine visits in the local area to develop a recommendation model based on the similarity of geospatial and personal traits of the elderly. The results of the study will be used to increase the activity of the community by combining a more effective analysis method with a practical recommendation program. | en_US |