A Study On The System Of Technology Project Similarity Detection

With the government’s growing investment to technology projects, the number and range of project application are both becoming larger than before, which bring harder work for review and assessment and make management more difficult. Many repeated projects have emerged. At present, document copy detection technique gets wide application in protecting intellectual property and optimizing search engine, but receives the least attention in technology projects management. The paper mainly proposes a similarity computation method based on fields according the technology projects knowledge representation, and develops a system that can efficiently and accurately find similar projects, provides warning to the projects reviewer and prevents the repeated projects effectively. The major research is as follows:Firstly, considering the representation of technology projects knowledge, a model that combined Vector Space Model with Matter-Element Model is proposed. The model gains keywords by using word segmentation for technology projects, then computes keyword weight by term frequently method.Secondly, according the project knowledge representation model, a similarity computation method based on project field is proposed. It sets the single field knowledge as the term frequency statistic unit. The algorithm adopts string hash matching, and calculates similarity among projects from each field similarity’s weighted average whose corresponding individual similarity is from vector cosine formula of technology projects.In the end, a similarity detection system of technology projects based on the above researches is developed. The system is formed with project knowledge base, project knowledge building module, similarity computation module, judgment and explanation module, and parallel computing task management module. Project knowledge building module mixes the knowledge model of projects to be examined and has been set up; similarity computation module calculates the similarity according to the knowledge model between projects; judgment and explanation module estimates the similarity relation on the basis of similarity, then the parallel computing task management module is used to do parallel processing to similarity computing module and judgment module. The technology projects similarity detection system has been applied to Zhejiang technology projects management system, and practical application shows the feasibility and effectiveness of the system that the paper proposed.