Research On Personalized Search Based On Context Awareness

With the rapid development of earth observation technology,the number of map products for disaster emergency has shown a trend of multiplication.The result of emergency mapping is to send disaster information to specific users in the form of map products.Whether users can quickly and effectively obtain disaster information directly determines the effectiveness of maps.Therefore,with rich in the map products nowadays,how to fast and precisely provide precise and personalized emergency map groups to support their decision-making is the challenge what the producers of map products currently facing.Emergency map products need to meet the needs of the users for high timeliness,namely different groups of users could use the relevant emergency map service system to quickly and effectively obtain the required map group information in the first time immediately after the disaster.However,the traditional emergency map service system mostly adopts the direct matching method of keywords as requested by the user for searching which lacks the personalized retrieval method for different types of users Therefore,taking the knowledge base of fire emergency map group as a case,this thesis proposes a personalized search method based on context awareness.Firstly,this thesis investigates the domain knowledge base of graph group based on fire emergency scene and the construction method of user situation model.Secondly,an ontology-based extended retrieval algorithm is proposed,and on this basis,how to add user context information in each retrieval process is studied,so as to realize personalized retrieval for different users.Finally,the retrieval algorithm is implemented through the system service platform.The main contents of this thesis include the following aspects:(1)Construction of domain knowledge base of fire emergency map group.Firstly,this thesis studies various stages of fire and their scene description methods,Secondly,this thesis refines the main concepts of domain ontology according to relevant specifications and plans,and finally constructs the ontology by using seven-step method,and formally represents and stores it in database.(2)Construction of user context model.Firstly,the relationship between user context and map group knowledge base is discussed.Secondly,the user context in disaster management is defined,and the corresponding context factors that need to be considered by users in the retrieval process are defined.Finally,all kinds of users in disasters are classified,and the database storage method of user context is designed.(3)Personalized search framework based on context awareness.Firstly,the user's retrieval statements are processed,keywords are extracted and their similarity is calculated.Secondly,an activation diffusion algorithm based on Word2 vec model is proposed for ontology query expansion.Finally,on the basis of the above research,how to combine the user context in the retrieval process is discussed,and a related application system is implemented to demonstrate the retrieval function.The personalized search method based on context awareness systematically elaborates the retrieval process at the theoretical level,and makes intensive research from the analysis of retrieval sentences to the ranking of results.At the method level,the ontology query expansion algorithm for ontology domain is proposed and the personalized search process is designed.At the application level,the prototype retrieval system of graph group is designed and implemented,and the retrieval framework proposed in this thesis is implemented and verified.