Multi-State Markov Reward Models For Academic Capacity Measures Of Universities

Measure academic capacity assessment is one of the important areas of innovation study in colleges and universities.University is an important body of the national scientific and technological innovation system. By research academic capacity measures of universities, On the one hand can meet the needs of Chinese economic and social development, on the other hand can also promote the development of its own universities.Academic capacity assessment should include two aspects:academic success and its application, academic stability support. In both theory and practice is a complex task, we achieved primarily through the following research study to measure academic capacity of universities.(1)In this paper, the development of both domestic and foreign universities rankings and academic capacities research was reviewed, also pointed out the shortcomings of the current evaluation system. And corresponding countermeasures were suggested. In a clear meaning on the basis of academic capacity,discussed the necessity of academic capacity measures of universities. The paper also proposed the selecting principle of evaluation indicator. In accordance with the principle of selecting indicators and AHP reasonable ability to build an indicator system for academic capacity measurements.(2)This paper proposed multi-state Markov reward models for academic capacity measures of universities.Greatest feature of the model is able to predict the trend of academic capacity,the result can be used as a basis for decision-making. According to the evaluation indicator system, processed and standardize sample data of two universities.The state is divided according to two-dimensional model, draw an acceptable state and unacceptable state set, and constructed state space successfully. By using different transition probabilities in calculation of multi-state Markov reward models for academic capacity measures of universities, predicted the reliability and the number of enter into unacceptable states of the two sample universities. Made the comparison of different initial state in the demonstration process.(3)Proposed two different ways to seek probabilities of state transition.One is based on the "bathtub curve" principle, the second is based on mathematical statistical calculations.Confirming the space of state and calculate the probability of transition were two prerequisites in using the Markov reward model. Through the empirical studies, the model had shown great practical value.