Hybridizing principles of TOPSIS with case-based reasoning for business failure prediction

Abstract

Case-based reasoning (CBR) solves many real-world problems under the assumption that similar observations have similar outputs. As an implementation of this assumption and inspired by the technique for order performance by the similarity to ideal solution (TOPSIS), this paper proposes a new type of multiple criteria CBR method for binary business failure prediction (BFP) with similarities to positive and negative ideal cases (SPNIC). Assuming that the binary prediction of business failure generates two results, i.e., failure and non-failure, we set the principle of this CBR forecasting method which is termed as SPNIC-based CBR as follows: new observations should have the same output as the positive or negative ideal case to which they are more similar. From the perspective of CBR, the SPNIC-based CBR forecasting method consists of R⁴ processes: retrieving positive and negative ideal cases, reusing solutions of ideal cases to forecast, retain cases, and reconstruct the case base. As a demonstration, we applied this method to forecast business failure in China with three data representations of a formerly collected dataset from normal economic environment and a representation of a recently collected dataset from financial crisis environment. The results indicate that this new CBR forecasting method can produce significantly better short-term discriminate capability than comparative methods, except for support vector machine, in normal economic environment; On the contrary, it cannot produce acceptable performance in financial crisis environment. Further topics about this method are discussed.

Introduction

One of the most important activities in management is performance measurement. Employers, governmental officials, investors, creditors, and bankers need effective tools to help them identify companies with good performance from those that perform badly. Investing or working in companies with potential developing capabilities will reduce risk. If a company goes bankruptcy or financial distress, then the company is in business failure. Research on business failure prediction (BFP) can provide tools for identifying companies with good performance [1], [2], [3], [4], [5], [6], [7], [8], [9]. Nowadays, companies and their suppliers and consumer companies have comprised supply chains to survive and make more money in global competition. Companies must therefore make their supply chains more competitive by cooperating with trustworthy companies with good performance. For companies, it is sensible to use BFP as a pre-diagnosis tool for supply chain trust diagnosis because companies with high performance should be trusted and employed in order to build a competitive supply chain with minimal risk. If it is predicted that a company will fail, this company does not deserve trust from other companies and should not be chosen as part of a supply chain unless it adopts measures to improve its performance. Some other criteria can be further used when making the final decision of supplier evaluation, e.g. quality, delivery, price/cost, manufacturing capability, service, management, technology, research and development, flexibility, reputation, relationship, risk, safety, and environment [10].

BFP can be used as a tool to check the credit risk of companies. The results and principles of such a tool must be easily interpretable. For example, an investor can employ BFP to find out whether or not a company deserves his or her investment. BFP can also be used by companies as a tool for supply chain trust diagnosis. Case-based reasoning (CBR) is an artificial intelligence approach which simulates the human problem-solving mechanism of solving new problems by recalling similar experiences. It is based on the assumption that similar observations have similar outputs. Due to its characteristics of ease of computation and ease of understanding, the chief decision-aiding technique for order performance by similarity to ideal solutions (TOPSIS) has been widely used in solving decision problems by functioning preferences of human beings. Deriving the concept of displaced ideal point from which the compromised solution would have the shortest distance, TOPSIS is a useful technique developed in the area of multi-criteria decision making. Under the assumption that the current problem may have several alternative solutions, this method ranks alternatives by comparing alternatives’ distances with the positive and negative ideal solutions, and then generating one preferred solution from alternatives. This preferred alternative is the one which is more similar to the positive ideal solution than other alternatives and is more dissimilar to the negative ideal solution than other alternatives. The inside mechanism is easily interpretable and explainable. Although TOPSIS is not designed to solve forecasting problems, comparison of distances with positive and negative solutions can be regarded as a specific implementation of the idea that similar observations have similar outputs. If an alternative is similar to the positive ideal solution, how well it can solve the current problem is correlated with how similar it is to the positive ideal solution. Inspired by the idea of TOPSIS, we attempt to investigate and apply a new type of CBR method to solve the problem of BFP.

By revising principles of TOPSIS under the assumption that similar observations have similar outputs, this research is devoted to proposing a new CBR forecasting method of BFP based on the similarities to positive and negative ideal cases (SPNIC). This method belongs to a type of multiple criteria CBR (MCCBR), which refers to a hybrid method from combination of multiple criteria decision-aiding techniques with CBR. This forecasting method retains characteristics of ease of interpretation and explanation, and serves as an effective alternative predictive tool of BFP. This paper is organized as follows. Section 2 introduces basic concepts of CBR and TOPSIS, with a discussion on significance of this research. Principles and procedures of the new CBR, i.e., SPNIC-based BFP method, are proposed in Section 3. The new forecasting method is applied to predict the business failure of Chinese companies with three data representations of a dataset from normal economic environment and a dataset from financial crisis environment in Section 4. Finally, conclusions and research limitations are discussed in Section 5.