build me a data mining project
Budget: $30 – $250 USD
project deadline is : 11/18/2022
project pay max : $135
project definition:
The manager of the bank wants to create an intelligent system that can segregate the people
into credit score brackets to reduce the manual efforts of checking their credit scores. The
system should have all necessary data required by the bank and other stakeholders.
Using the example of a person’s credit-related information, we will build a machine learning
model that can determine whether an individual is a credit risk or not. We will be using a
supervised method where we train the algorithm with real-world training data and then use it
to predict the outcome for new instances.
Dataset: (Description)
In the train and test data set, there are 28 columns with attributes like ID, Customer ID, Month,
Name, Age, and more. By using PCA and correlation for feature reduction can gain insights from
the data if we're looking for patterns and variations.
Source of Dataset:
https://www.kaggle.com/datasets/parisrohan/credit-score-classification
You must use only above dataset
The following are the models selected by us
1) EDA (Data Cleaning, Handling Missing data)
2) CART
3) KNN with Grid search,
4) Random Forest
5) Bagging Decision Tree
6) Logistic Regression with Grid search
7) Support Vector Machines (SVM) with Grid search
8) AdaBoost Classifier
9) XGBoost
NOTE : Classifications no 6,7,8,9 you can change if you want (You replace with other Classifications) we need total 8 classifications and we cannot consider this step as count classification EDA (Data Cleaning, Handling Missing data)
this is your concussion
Model metrics and Evaluation:
For balanced datasets, accuracy and precision are good measures to evaluate classification
models, but if the data is unbalanced, alternative approaches, including ROC/AUC, perform
better
here same requirements
1. Choose a dataset which has 2 class labels/ Target variable(Y/N, 0/1).
2. The number of input features should atleast be 12 to 15.(excluding output class label column)
project pay max : $135
project definition:
The manager of the bank wants to create an intelligent system that can segregate the people
into credit score brackets to reduce the manual efforts of checking their credit scores. The
system should have all necessary data required by the bank and other stakeholders.
Using the example of a person’s credit-related information, we will build a machine learning
model that can determine whether an individual is a credit risk or not. We will be using a
supervised method where we train the algorithm with real-world training data and then use it
to predict the outcome for new instances.
Dataset: (Description)
In the train and test data set, there are 28 columns with attributes like ID, Customer ID, Month,
Name, Age, and more. By using PCA and correlation for feature reduction can gain insights from
the data if we're looking for patterns and variations.
Source of Dataset:
https://www.kaggle.com/datasets/parisrohan/credit-score-classification
You must use only above dataset
The following are the models selected by us
1) EDA (Data Cleaning, Handling Missing data)
2) CART
3) KNN with Grid search,
4) Random Forest
5) Bagging Decision Tree
6) Logistic Regression with Grid search
7) Support Vector Machines (SVM) with Grid search
8) AdaBoost Classifier
9) XGBoost
NOTE : Classifications no 6,7,8,9 you can change if you want (You replace with other Classifications) we need total 8 classifications and we cannot consider this step as count classification EDA (Data Cleaning, Handling Missing data)
this is your concussion
Model metrics and Evaluation:
For balanced datasets, accuracy and precision are good measures to evaluate classification
models, but if the data is unbalanced, alternative approaches, including ROC/AUC, perform
better
here same requirements
1. Choose a dataset which has 2 class labels/ Target variable(Y/N, 0/1).
2. The number of input features should atleast be 12 to 15.(excluding output class label column)