User Loneliness Recognition And Pattern Analysis Research Based On Smartphone

Loneliness,originated from the field of psychology,has been paid more and more attention in today's daily life.Especially with the development of mobile Internet technology,smartphones is no longer a simple communication tool,but a set of games,cameras,Internet access,social networking,information access and other tools as one of the multi-functional personal handheld terminal.As people rely more and more on the Internet and smartphone,"bow family" in our daily lives everywhere.Studies have shown that there is a strong relationship between mental health and smartphones.On the one hand,smartphones have changed people's psychological conditions,on the other hand,people's psychological conditions also affect the use of smartphones.We can collect phone data from smartphone in a ubiquitous way,analyze the relationship between the data and loneliness offline,and establish the classification models to judge the loneliness of user in real time.Therefore,the loneliness recognition based on smartphone data has a very important research value and practical significance.How to use the data collected from smartphones to identify the loneliness of users is a new research field in academia.In this paper,we use machine learning methods to study the state of loneliness,design and implement a prototype system,which can identify the loneliness of users in real time.The main work of this paper including the following 3 aspects:(1)In order to find out the relationship between loneliness and smartphone,we regard the problem of loneliness recognition as a binary classification problem.Through a series of data processing and analysis,the machine learning methods are used to classify the data sets.The result shows that the highest accuracy rate is 69.5%,which confirms the relationship between loneliness and mobile phone.(2)In order to determine the behavior patterns of the lonely people,we divide the whole population into two categories according to the loneliness: loneliness and loneliness,and then analyze and compare the behaviors and patterns of the two groups.We summarize the behavior patterns of lonely people and their characteristics,the result shows that there is a big difference between the two groups.(3)In order to realize the real-time recognition of the lonely state,we design and implement the user-based state identification framework based on cloud,construct the incremental classification model by incremental learning,and establish a variety of classification models,compare the differences between models.According to the actual situation of each user to automatically identify the user's lonely state,through the recruitment of volunteers to experiment,and finally up to 76.83% accuracy,verify the effectiveness of the prototype system.