Abstract
This study investigates the dynamic process of the company’s financial status over years and proposes to design financial path using fuzzy c-means (FCM) approach. FCM is firstly used to quantize company financial paths of behavior over consecutive years. Financial paths are deployed to depict different patterns of failure and to understand the dynamics of financial failure. Then financial failure prediction model is built based on the proposed financial path (FP) approach. Empirical experiment is carried out with data samples of Chinese listed companies. Through analyzing financial path, it is found that there are mainly four patterns of process to terminal failure. Besides, in order to validate the prediction performance of the financial path prediction model, four prevalent financial failure prediction models, logistic regression (LR), support vector machine (SVM), decision tree (DT), and neural network (NN), are deployed to compare with the proposed model respectively. Experimental results show that FP has significantly better financial failure performance than other four models in terms of accuracy, type I error, and type II error. Therefore, the financial path monitors the change of financial status and acts as a prediction tool for financial failure prediction, which is a significant determinant of the financial success. Managers can recognize the financial failure signal in advance and understand their evolution trend on the future. In addition, the financial path prediction model is also an effective supplement to the research field of financial failure analysis and prediction.
Highlights
Financial failure prediction model is built based on the proposed financial path (FP) approach
A cial path monitors the change of financial status and acts as a prediction tool for financial failure prediction, which is a significant determinant of the financial success
D and analyze the dynamics of financial failure. It is for this reason that we study this issue –financial path analysis –from a quantitative point of view
Summary
Most financial failure models are single-period models. These models are designed using financial variables collected at time t, and their prediction performance is measure at time t+1, t+2, or t+3. Almost all classic financial failure prediction models are subjected to the problem related to using only. The transverse data samples (i.e. one annual account) to build the prediction models. The use of transverse financial data is because they neglect the important factor of time dimension in building models. Financial failure prediction models can be divided into two categories based on the idea of modelling techniques, they are snapshot model and historical model respectively. Snapshot model ignore the time factor when designing models, whereas historical model introduces time as a dimension when building prediction models
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call