INTELLIGENT SERVICE PUSH METHOD BASED ON ACTIVE GEOGRAPHIC PERCEPTION

In view of the lack of consideration of user behavior motives in traditional personalized precision service systems, the accuracy of service content is not high.In order to solve this problem, research on personalized accurate service push method based on active geographic perception. By constructing a geographic feature information model, get the characteristics of the user's destination in real time, and then infer the user's behavioral motivation. Focusing on active geographic awareness technology and personalized precision service methods, the concept, principle, process and key technologies of active geographic sensing are studied, determined the main research content of active geographic perception and the relationship. Then analyze and discuss the construction method of active geographic awareness architecture, developed a geographic feature content system and studied its extraction and weight calculation methods. By the way, according to the characteristics of active geo-sensing, an active awareness API conforming to high efficiency and real-time is designed. Then explored the personalized accurate service push method based on active geographic perception,designed three processes of geographic awareness, service retrieval and service push, a service retrieval and delivery method is proposed. Finally, a personalized precise service system based on active geographical perception is designed. By adding geographic features to the personalized precision service, it can make up for the lack of service personalization and lack of precision caused by ignoring user motivation, which provides a new idea for more accurate and personalized service push. * Jingwen Li Lijw@glut.edu.cn


INTRODUCTION
Under the multiple influences of big data, cloud computing, mobile Internet and in-depth learning,Enhancing the wisdom of service, increasing the individualization of service content and improving the accuracy of service have become three main problems in service industry research [1] . Around the above three issues,many scholars have carried out in-depth research and put forward a large number of theoretical and technical support.However, most of the current research only considers the user and project dimensions, which leads to the generalization and solidification of the recommended algorithm. The reason for this problem is that the user's activities are not only affected by personal preferences, but also related to the user's situation. Situation reflects the user's behavioral motivation,the unified intelligent service can be universally applied, but because of the less motivation for considering user behavior, when the user has a subjective purpose, the accuracy of the service will be deviated. In order to solve this problem, this paper studies the user motivation inference method based on active geographic perception, which narrows the user's preference to the time level, and then improves the accuracy of intelligent service.

The Meaning of Active Geographic Perception
Human behavior is often purposeful, and the purpose is related to the subjective thoughts of the individual and the characteristics of the destination. For example, if a person who likes food but does not have significant digital hobbies goes to the digital shopping area, then his potential purpose should be to buy digital products,But traditional recommendation systems recommend gourmet products to users through their preference characteristics [2] ,this will lead to the generalization of recommended content to a certain extent, reducing the accuracy of recommendations [3] . In order to solve this problem, we need to combine user characteristics and situational features in the service recommendation. In this paper, the user characteristics obtained through a large amount of data analysis and mining are called universal features, and the user characteristics (behavior preference) in a specific situation at a certain moment is called the context feature. Active geographic perception is an intelligent service recommendation technology that takes the user's preference characteristics from the universal level to the context level. Active geographic perception relies on the multi-dimensional geographic spatio-temporal data model. The destination features are indexed by time, and the spatial and attribute dimensions are collaboratively described. They are tagged at the natural, social and commercial levels, and their characteristics are abstracted to feature. The text reconstructs the destination context, and combines the user preference characteristics to infer the behavioral motives of the user at a specific place at a specific time, and finally calculates the user's situational characteristics. Active geo-aware technology can reduce the user's preference to the context level by reconstructing the situation, which can solve the problem of generalizing and solidifying push content in some cases, and improve the accuracy of intelligent services. High-precision positioning technology is the trigger basis for active geographic sensing, high-precision real-time positioning mainly solves location acquisition and initiative realization in active geo-sensing technology. Initiative is achieved based on changes in location, but since the user is moving, the position is changing in real time, not all position changes will trigger context reconstruction. In this paper, by combining the destination range, it is judged whether it is a valid position trigger by setting the position change value, thereby improving the service efficiency of the system.

Active Geographic Perception Process
Active geographic perception service flow is divided into two phases. The first stage is to acquire the longitude and latitude of mobile devices in real time according to positioning technology, and do inverse address resolution to the longitude and latitude to get the destination text, the second stage uses the destination as a key to retrieve the geographic feature information database, obtain the natural, social and commercial geographic feature information of the location, and infer the user's behavioral motives based on the user characteristics. The flow chart is as follows: Figure 1 Active geographic awareness flow chart The core content of active geographic awareness is to build a geographic feature database. This paper constructs a geographic feature database from three levels of business, nature and society.

Feature classification
This article divides geographic features into three categories: natural, social, and commercial. The specific classification is as follows: (1) Natural characteristics Natural characteristics are the characteristics of natural environment in a region. Because of the temporal nature of natural characteristics, they can be divided into dynamic and static categories according to their frequency of change.  Table   3.2 Feature extraction Natural and commercial features are characterized by stability and wide range. These features are obtained through a large Q&A platform. Commercial data ranges are relatively small The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W10, 2020 International Conference on Geomatics in the Big Data Era (ICGBD), 15-17 November 2019, Guilin, Guangxi, China (blocks), and commercial data has a determined latitude and longitude, so this type of data is obtained from POI data [9] .
(1) Natural and social feature extraction Dynamic features are acquired in real time through an open API interface, and static and social features are implemented using network data capture and manual intervention auditing. The network crawling features mainly include four steps: feature text acquisition, Chinese word segmentation, feature extraction and manual review. The flow chart of feature extraction is as follows: Figure 2 Feature extraction process (2) Commercial feature extraction The commercial feature is calculated based on the POI data, and the calculation is based on the latitude and longitude of the POI. Calculate the proportion of different types of POIs in a specific area based on latitude and longitude,this ratio is the commercial characteristics of the area.

USER BEHAVIOR MOTIVATION INFERENCE METHOD BASED ON ACTIVE GEOGRAPHIC PERCEPTION
The ultimate goal of active geographic awareness is to infer the user's behavioral motivation by reconstructing the situation, the process is: Firstly, the geographical feature information of the user's location is obtained, which includes the geographic feature tag and the feature weight. Secondly, the user's preference information is retrieved. Finally, based on the geographic feature and user preference, the association weight is used to obtain the feature weight of the combination. According to the weight, the user behavior motivation is inferred, and the user does not push the more accurate and personalized service. User motivation inference flow chart is as follows: Figure 5 User motivation inference process The most central part of inferring user motivation is the association analysis of geographic features and user preferences, The process of its analysis is as follows: