Comment
Author: Admin | 2025-04-28
Non-performed loans. Dataset includes 916 records of accepted loan borrowers and 584 records of not accepted loan borrowers. 2. Data Pre-processing Data Pre-processing is a major step in the Machine Learning process. In this step, data is cleaned, transformed and encoded to make an algorithm to easily interpret the features of the dataset. Real-world data is frequently inconsistent, incomplete, erroneous and lacking in specific behaviours or trends. For resolving such problems, pre-processing is used in the Machine Learning Model development process (Pandey, 2019). Data pre-processing steps that have been used in this research are as follows. Missing values were replaced by mean values of each column. Discretization was done to the two numerical attributes called customer Id (CID) and amount. CID value was divided into 0-1500 range and amount attribute was divided into 0-12 range. 3. Feature Selection In this research study, initially there were 17 attributes and among those attributes, most correlated attributes have been selected using correlation matrix by considering the relationship between independent variables and the dependent variable. Correlation matrix determines correlation between all attributes, and it can produce a weights vector based on these correlations (Kumar, 2018). Correlation is a statistical technique that can show whether and how strongly pairs of attributes are related. Correlation Matrix has been used to identify the special parameters which have strong positive relationships towards the label of the data set. According to the results received from the correlation matrix, out of 17 parameters there were 12 special parameters which have a strong positive relationship towards the label. Considering the correlated matrix results, following features were selected for model implementation. CID - Customer identification number given to each customer Amount - Loan amount (values between 500,000 to 100,000,000) Facility Type - Loan type (Permanent Over Draft(
Add Comment