The Deign And Implementation Of The Employee Quality Assessment Based On SVM In HRMS

With the rapid development of economy and the sharply rise of knowledge economy, we gradually recognize that human are the main knowledge carrier and user. Talent competition becomes the key impact in the competition among companies. Thus, more and more companies intend to pay more attention to their human resource management and talent selection.Taking human resource management of enterprises and companies as research background, this thesis designs and implements a human resource management system based on B/S to realize information and the practice proves the HRMS can improve the office efficiency observably. The system adopts B/S three layer architecture, which makes the system more suitable for distributed office environment. At the same time, the system based B/S architecture also can improve information transmission speed in inter company. Besides, we adopt MVC 3 layers development framework, which increases the robustness of the software, and makes it more convenient for software maintenance and extension.In the system requirements analysis phase, we portray the needs of human resource management, including the overall requirements, functional requirements and non-functional requirements based on UML. In functional requirement modeling, use case analysis is adapted to interpret the organization management, job management, and staff management; in addition, the paper describes the non-functional requirements of the system through three aspects: reliability, ease of use and safety. In the system design stage, through physical and logical architecture, we describe the overall architecture design. In detail functional design process, the 3 business models are designed corresponding to the requirement analysis. Besides, we also illustrate the relational database model of the HRMS thought the ER diagrams and information tables.Based on the traditional HRMS, this paper also focuses on the talent selection. We design an algorithm based on SVM for talent review in order to support the enterprise manager valuable data for human resource management. Meanwhile, this function also can make the enterprise inner management fairer. In the SVM design process, we model a perfect feature vector related to staff’s quality and build dataset with the supervising of domain experts. Then, based on binary tree we train multiple SVM binary classifiers. At last, we can resolve the multi-class problem for talent review. However, with the rapid development of economy, the standard of talent would change. Thus, the SVM classifiers should be adaptive with the added mass data.Finally, this thesis introduces the implement of key function models and the system deployment. Besides, we also explain the system experiments and analysis the experiment results.