Collection Scoring Models Development and Research Based on the Deductor Analytical Platform

Abstract

This article solves the problem of collection scoring models constructing and researching. The relevance of solving this problem on the intelligent modeling technologies basis: decision trees, logistic regression and neural networks is noted. The initial data for the models was a set of 14 columns and 5779 rows. The models construction was performed in Deductor platform. Each model was tested on the set of 462 records. For all models, the corresponding classification matrix were constructed and the1st and 2nd kind errors were calculated, as well as the general error of the models. In terms of minimizing these errors, logistic regression showed the worst results, and the neural network showed the best. In addition, the constructed models effectiveness was evaluated according to «income» and «time» criteria. By the time costs the logistic regression model exceeds other models. However, in terms of income the neural network model was the best. Thus, the results showed that in order to minimize the time spent on work with debtors it is advisable to use a logistic model. However, to maximize profits and minimize classification errors, it is appropriate to use a neural network model. This indicates its effectiveness and practical use possibility in intelligent scoring systems.

Automatic Scoring Model of English Interpretation Based on Semantic Scoring

“…Gillier et al proposed that, at major international conferences such as medicine and law, the audience's views on the interpretation of quality assessment were summarized and case studies were conducted, and it was found that there were diferences in the interpretation of the quality assessment standards between audiences and translators. Ismagilov and I interviewed members of the International Conference of Translators, and the study found that audiences and translators place a high value on the accuracy and clarity of translated content [14].…”

Trajectory classification-based car loan collection scoring model using a hierarchical CNN-LSTM method

“…To compare our method with existing collection scoring methods, we collect registration information and repayment records which are used in References 3,4 as many as possible for all the clients in the dataset. A description collected variables is presented in Table 3.…”

“…Then machine learning algorithms like Logistic regression, SVM and Decision Trees are embraced to make a prediction based on these data. In Reference 3 , 12 variables are used (age, number of payments before the delay, etc.) to train a logistic regression model to segment overdue clients into groups.…”