Bio-inspired optimization of an incrementally updated fuzzy investment decision support system
Lars Kruger, Matthias Walter, Jayesh Jani · 2013
The system we propose allows the classification of future performances of high-technology venture investments on the basis of very limited information. Our system thus helps investors to decide whether to invest in a young High-Technology Venture (HTV) or not. In order to cope with uncertain data we apply a Fuzzy Rule based Classifier. As we want to attain an objective and clear decision making process we implement a learning algorithm that learns rules from given real-world examples. The availability of data on early-stage investments is typically limited. For this reason we equipped our system with a bootstrapping mechanism which multiplies the number of examples without changing the inherent quality or structure of the examples. To enhance the performance of the IDSS we apply a specifically designed Particle Swarm Optimization algorithm (PSO). We show the efficacy of this approach by comparing the classification power and other metrics of the PSO-optimized system with the corresponding characteristics of the original IDSS.