Enhancing credit scoring with alternative data
نویسندگان
چکیده
Hundreds of millions people in low-income economies do not have a credit or bank account because they insufficient history for score to be ascribed them. In this paper we evaluate the predictive accuracy models using alternative data, that may used instead history, predict risk new account. Without type data is typically available demographic data. We show model contains email usage and psychometric variables, as well can give greater than uses only sufficiently high when conventional unavailable. The same applies if merely included together with However, different randomly selected training: test sample splits wide range accuracies. second part paper, two datasets include predictor, compare performances machine learning statistical classifiers. find some classifiers applied these predictors accurate predictions variables no other available.
منابع مشابه
CFPB Insights On Alternative Data Use In Credit Scoring
The growing utilization of alternative data in credit decisions and scoring models continues to attract the attention of regulators. Although clear and practical guidance from regulators remains elusive, regulator pronouncements provide some clues and takeaways for market participants. Recently, for example, the Consumer Financial Protection Bureau issued a request for information (RFI) regardi...
متن کاملEnhancing credit scoring model performance by a hybrid scoring matrix
Competition of the consumer credit market in Taiwan has become severe recently. Therefore, most financial institutions actively develop credit scoring models based on assessments of the credit approval of new customers and the credit risk management of existing customers. This study uses a genetic algorithm for feature selection and decision trees for customer segmentation. Moreover, it utilize...
متن کاملConsumer credit scoring models with limited data
In this paper we design the neural network consumer credit scoring models for financial institutions where data usually used in previous research are not available. We use extensive primarily accounting data set on transactions and account balances of clients available in each financial institution. As many of these numerous variables are correlated and have very questionable information conten...
متن کاملCredit Scoring and Data Mining
Credit scoring is the use of predictive modelling techniques to support decision making in lending. It is a field of immense practical value that also supports a modest amount of academic research. Interestingly, the academic research tends not to be put into practice. This is not a result of insularity and arrogance on the part of the practitioners, but rather, of the practitioners having a be...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Expert Systems With Applications
سال: 2021
ISSN: ['1873-6793', '0957-4174']
DOI: https://doi.org/10.1016/j.eswa.2020.113766