Artificial Neural Network Model In Project Investment Risk Analysis

This paper focuses on the Artificial Neural Network Model in Project Investment Risk Analysis. Starting with the phases of risk analysis, a systematic method is introduced including inputting the historic data and expert assessment into the risk identification unit, and building 3-tier feed-forward neural network model and make adjustment to the network weights for training so that finally a risk estimation coefficient is obtained. Through all this, a new idea on modern algorithm's application to the investment risk analysis is introduced and is foreseen to have a good prospect for quantified and scientific project risk analysis.A software based on ANN in project investment risk analysis model is developed in this paper, which can be identified as an integration of three units: Risk Identification Unit, Risk analysis Unit based on ANN model and Risk Evaluation Unit. Java and Delphi development platform are adopted in construction of the application so that good cross-platform, translatability and extensibility is implemented in the software.