Global Journal of Management and Business Research, A: Administration and Management, Volume 22 Issue 1

Table 1: Company performance assessment criteria (source: [3]) ① ② ③ ④ ⑤ ⑥ ⑦ ⑧ Operating CF ＋ ＋ ＋ ＋ ― ― ― ― Investing CF ＋ ― ＋ ― ＋ ― ＋ ― Financing CF ＋ ― ― ＋ ＋ ＋ ― ― Ishikawa et al. [4], Mizoguchi et al. [5], and Jiang [6] employed a discriminant analysis and proposed predictive models for binary bankruptcy/non- bankruptcy outcomes based on machine learning using their respective datasets described in the “Data used” column of Table 2. In addition, Jidaisho et al. [7] used logistic regression to analyze their data, also shown in Table 2, for machine learning and proposed a prediction model for a binary bankruptcy/non-bankruptcy assessment. Masuyama[8] also analyzed the financial statements of bankrupt companies by chronologically organizing their data, as described in Table 2. They also drew on surveys administered by the Small Business Institute Japan and management improvement plans of individual companies to compare the actions taken to avoid bankruptcy, based on which they attempted to conduct a bankruptcy prediction. Finally, Saigo et al. [9] developed a model to evaluate companies by applying the discounted cash flow (DCF) formula to the free cash flow. DCF is a valuation method used to estimate the corporate value at certain discount rates based on the future cash flow expected from a business. Saigo et al. specifically addressed SMEs and micro-businesses and discussed measures to improve their corporate value based on the DCF. For example, they described “cutting unnecessary investments” and “optimizing the equity structure” as financial optimization measures to attain the lowest discount rate, which is one of the components of DCF, and “enhancing the business efficiency” and “investing in profitable businesses (business portfolio optimization)” to maximize the corporate value. Most of the studies above applied machine learning to their respective financial data and attempted to attain binary outcomes between bankruptcy and non- bankruptcy prediction results. Whereas Masuyama and Jiang both went further and considered those factors responsible for bankruptcy, Masuyama only discussed the factors drawing on some case studies, and the latter employed another method to analyze the detrimental factors after using the discriminant analysis for the binary prediction. They fall short of integrating a series of analyses into a single automated process. Therefore, we developed a model that facilitates not only the prediction but also the identification of financial variables that may contribute to the bankruptcy of a company. Table 2: Examples of existing studies on bankruptcy prediction models concerning Japanese companies Study title Prediction model Analytical method Data used Yosuke Kono et al.: Discussion on the Possibility of Predicting Corporate Bankruptcy [2] Plotted the mean values of the data organized by fiscal years and compared between the sample data (of bankrupt companies) and the data taken from five listed companies 1. Current ratio 2. Equity ratio 3. Quick ratio 4. Monthly sales to current capital ratio 5. Fixed assets to fixed liability ratio 6. Fixed ratio 7. Interest bearing debt to monthly sales ratio 8. Return on assets 9. Equity ratio Ayaka Okubo: Study on Black-in Bankruptcy Mechanism through Financial Statements Focused on Cash Flow Statement [3] Developed 8 patterns of corporate financial states according to the combinations of positive and negative values each for operating, investing, and financing cash flows. 1. Operating cash flow 2. Investing cash flow 3. Financing cash flow Masaru Ishikawa & Ngai Chung Sze: A Study of Corporate Bankruptcies Based on the Cash Flow Information [4] Discriminant analysis 1. Operating cash flow 2. Investing cash flow 3. Financing cash flow 4. Operating cash flow margin: Operating CF/Sales revenue 5. Corporate CF to sales ratio: (Operating CF + Investing CF)/Sales revenue 6. Total assets to operating CF ratio: 10 Global Journal of Management and Business Research Volume XXII Issue I Version I Year 2022 ( ) A © 2022 Global Journals A Study on Machine Learning Prediction Model for Company Bankruptcy using Features in Time Series Financial Data