| Summary: | Model of neuro-expert system was developed. The main idea is to integrate both artificial intelligence methods such as neural network and expert system in a business strategy or investation. The neural-network is used to solve numerically the mathematical model developed based on the data collected from Jakarta Trade Center and the mechanism of acquiring the investation. The objective is to provide certain decision and recommendation to buy, hold or sell the insvestation based on the analytical result. Two aspects are integrated in this model such as fundamental aspects consisting of macro-economy, industry and enterprises conditions and also the technical aspects consists of the price of investation, knowledge bases collected from many experts and also the algorithm of both neural network and decision searching. Backpropagation algorithm with ten hidden-layer and Levenverg-Marquartd method are used in structuring the model. The data used in testing the model is collected from PT. Astra International Tbk. during 225 days of trade. The results shows that the enterpirse is recommended to sell the investation in the next 40 days and this decision is expected to gain the benefit of Rp. L075 per exemplar excluded the dividen.
Keywords : integration model, neuro-expert system, decision support, investation.
|
|---|