DESIGN AND EVALUATION OF A PROTOTYPE FUZZY-LOGIC-BASED KNOWLEDGE SYSTEM FOR FINANCIAL RATIO ANALYSIS

Development and validation of a knowledge system to support financial analysis was the goal of this research. Financial analysis was conceptualized as a problem solving process. Problems were defined to be aspects of financial performance and condition such as profitability, liquidity, solvency, etc. which concern the business enterprise. Financial ratios provide symptomatic evidence of problem existence and clues about possible causes. A financial problem solving model was developed to serve as a basis for a knowledge system.;FANFARE(s) liquidity assessments were validated against the standard of human expert judges. Nonparametric statistical tests (Kendall's Tau) indicated a strong ordinal association between FANFARE(s)' and judges liquidity assessments. FANFARE(s) generated cause possibility responses were validated against controlled simulation data.;FANFARE(s) (Financial Analysis by a Fuzzy Approximate Reasoning Expert (System)) is a prototype knowledge system based on the problem solving model. Production rules compose the knowledge base and deductive inference is by fuzzy logic. Liquidity-illiquidity is the modeled problem domain. Receivables and inventory management and excessive fixed assets are potential causes of illiquidity which FANFARE(s) is programmed to identify. FANFARE(s) output consists of a linguistic value of liquidity and numeric possibility ratings of the three possible causes of illiquidity.