Exploratory data analysis is what I did throughout this stage of the process. And so I began by gaining an understanding of the data head, the data tail, the data shape, the data size, the data columns, the uniqueness of the data, and the type of variable that was contained within the data, not to mention the data types.
In this stage of cleaning the data, I remove any undesirable values and variables from the data set as well as any irregularities that may exist within it. This is because irregularities of this kind may have a negative impact on the findings we obtain. I checked for missing or null values by using the code "data.isnull().sum()," which sums up the data. luckily my data had no null values.
In the generation of the logistic regression, I got this result by executing the following line of code: [data.drop(['column name1', 'column name2'], axis=1, inplace=True)] in order to exclude columns from our logistic regression analysis that were found to be insignificant. I checked for liars and drew cutter graphs using the data set that was provided by the bank staff. After that, I performed a relationship analysis on the data by plotting a pair plot and a co-relation hit map. I got rid of everything that wasn't necessary from the data set. The next thing that needed to be done was to separate the data into test and training sets. Logistic regression was imported by utilizing the following piece of code: [from sklearn.linear model import LogisticRegression logreg is an instance of LogisticRegression, and logreg.fit(X train, y train) and y pred is an instance of logreg.predict(X test) are both functions. after that, the accuracy of the prediction and the raw confusion matrix were examined.