Objectives. The aim of the study is to examine whether and how artificial intelligence (AI) may facilitate the joint comprehension of corporate distress and corporate legality. The main subjects of investigation are both represented by the valuation of company’s distress and by the legality rating (LR), which is a measure of the company’s degree of legality). LR’s adoption allows firms to benefit from some advantages when accessing to credit. For this reason, LR is related to the company’s creditworthiness, and by consequence, to the company’s distress. Methodology. The dataset is composed by companies in possession of legality rating. AI is used as methodological approach. Decision trees allow to automatically identify combination of variables from the dataset that explains the two target variables, zone of discrimination and cut off, according to a different perspective, that is not considered by Z’ score. Findings. AI allows to identify a new “basket” of variables, different from those employed by the Altman’s Z’ score, that determine the company’s distress. The experiments test the “ability” of the algorithm to identify a combination of variables to predict the target. It is possible to analyze in which way these variables get along with each other in order to produce with accuracy the correct identification of the target variable. Research limits. The methodology needs to be adapted determining plausible interval for the variables identified by the decision trees. The dimensionality of the dataset can benefit from resampling the variables for the proposed methodology which, at the state of the art, suffer from problems of skewness. Practical implications. The AI methodology is able to process large amounts of records within a given dataset, so allowing to test the effectiveness of LR in the assessment of creditworthiness. Originality of the study. The recognition and composition of the new variables can be interpreted as a tool to strengthen the comprehension of company’s distress.
Company’s distress and legality under the magnifying glass of artificial intelligence: the contribution of decision trees to identify best practices / Barile, Sergio; Buzzi, Irene; D'Avanzo, Ernesto. - (2020), pp. 13-34. (Intervento presentato al convegno GRAND CHALLENGES: Companies and Universities working for a better society tenutosi a Pisa nel September 7-8, 2020).
Company’s distress and legality under the magnifying glass of artificial intelligence: the contribution of decision trees to identify best practices
Ernesto D'Avanzo
2020-01-01
Abstract
Objectives. The aim of the study is to examine whether and how artificial intelligence (AI) may facilitate the joint comprehension of corporate distress and corporate legality. The main subjects of investigation are both represented by the valuation of company’s distress and by the legality rating (LR), which is a measure of the company’s degree of legality). LR’s adoption allows firms to benefit from some advantages when accessing to credit. For this reason, LR is related to the company’s creditworthiness, and by consequence, to the company’s distress. Methodology. The dataset is composed by companies in possession of legality rating. AI is used as methodological approach. Decision trees allow to automatically identify combination of variables from the dataset that explains the two target variables, zone of discrimination and cut off, according to a different perspective, that is not considered by Z’ score. Findings. AI allows to identify a new “basket” of variables, different from those employed by the Altman’s Z’ score, that determine the company’s distress. The experiments test the “ability” of the algorithm to identify a combination of variables to predict the target. It is possible to analyze in which way these variables get along with each other in order to produce with accuracy the correct identification of the target variable. Research limits. The methodology needs to be adapted determining plausible interval for the variables identified by the decision trees. The dimensionality of the dataset can benefit from resampling the variables for the proposed methodology which, at the state of the art, suffer from problems of skewness. Practical implications. The AI methodology is able to process large amounts of records within a given dataset, so allowing to test the effectiveness of LR in the assessment of creditworthiness. Originality of the study. The recognition and composition of the new variables can be interpreted as a tool to strengthen the comprehension of company’s distress.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione