Performance evaluation of innovative enterprises based on Neural network-Kalman filter model

Building an innovative country is an important national strategy and developing a number of innovative enterprises is the key point to build an innovative country. How to evaluate the innovative enterprises performance scientifically has become the research hot spot. Based on deep analysis of performance evaluating status and characteristics of innovative enterprises, some indicators are selected and designed in the paper by using the generalized minimum variance analysis, which include both financial indicators that embody the profitability, operating ability, debt paying ability, development ability and non-financial indicators that reflect the technology innovating, business process, knowledge management and external stakeholders´ interest. Innovative enterprises performance evaluating model was established by using Neural network-Kalman filter model with 60 innovative enterprises in electonics industry as samples, model value were fit high with the actual performance by empirical tests. The case analysis demonstrates that the performance could be evaluated by appling evaluation model and weak indicators causing low performance could be found separately. Some weak indicators were sorted according to influence degree and some countermeasures to improve innovative enterprises performance were proposed based on analying main weak indicators.