Abstract: One of the main features to invest in stock exchange companies is their financial
performance. On the other hand, conventional evaluation methods such as data envelopment
analysis are not only a retrospective process, but are also a process, which are incomplete
and ineffective approaches to evaluate the companies in the future. To remove this problem,
it is required to plan an expert system for evaluating organizations when the online data are
received from stock exchange market. This paper deals with an approach for predicting the
online financial performance of companies when data are received in different time’s intervals.
The proposed approach is based on integrating fuzzy C-means (FCM), data envelopment
analysis (DEA) and artificial neural network (ANN). The classical FCM method is unable to
update the number of clusters and their members when the data are changed or the new data
are received. Hence, this method is developed in order to make dynamic features for the
number of clusters and clusters members in classical FCM. Then, DEA is used to evaluate
DMUs by using financial ratios to provide targets in neural network. Finally, the designed
network is trained and prepared for predicting companies’ future performance. The data on
Tehran Stock Market companies for six consecutive years (2007–2012) are used to show the
abilities of the proposed approach.