Salary_Prediction_with_DecisionTree

This repository contains Python code for analyzing salary data and building a Decision Tree Regression model for predicting total pay based on various features. The code includes data preprocessing steps, handling missing values, and using scikit-learn for machine learning. Explore the code to understand how to predict salaries with Decision Trees.

Table of Contents

• Introduction
• Installation
• Usage
• Data
• Machine Learning

Introduction

In this project, we aim to predict the total pay of employees based on different attributes, including job titles, years of experience, and more. We use a Decision Tree Regression model for this purpose.

Installation

To run the code in this repository, you need Python and the following libraries:

• pandas
• scikit-learn

You can install these libraries using pip:

pip install pandas scikit-learn

Usage

1. Clone this repository to your local machine:

git clone <https://github.com/prashantkparth/Salary_Prediction_with_DecisionTree>

2. Navigate to the project directory:

cd Salary_Prediction_with_DecisionTree

3. Run the Jupyter Notebook or Python script to analyze the salary data and train the Decision Tree Regression model.

Data

We used the "Salaries.csv" dataset for this project. The dataset contains information about employee salaries, job titles, and more. Make sure to update the data source if you want to use a different dataset.

Machine Learning

We utilized scikit-learn to build a Decision Tree Regression model. The code includes data preprocessing steps, handling missing values, and encoding categorical features.

Example code for training the model and making predictions

from sklearn.tree import DecisionTreeRegressor

Create and train the Decision Tree Regression model

model = DecisionTreeRegressor() model.fit(inputs_new, target)

Make predictions

predictions = model.predict([[3]])

Feel free to explore the code in the Jupyter Notebook or Python script for more details.

Code Explanation

Certainly, I'll explain each part of the provided code step by step:

import pandas as pd

salaries = pd.read_csv(r"C:\Users\prash\machine learning\Salaries.csv")

This code snippet imports the Pandas library and uses it to read a CSV file named "Salaries.csv" located at the specified file path. It loads the data into a Pandas DataFrame named salaries.

salaries = salaries[~(salaries[salaries.columns[7:]] == 0).any(axis=1)]
salaries = salaries[~(salaries[salaries.columns[7:]] == -618.13).any(axis=1)]

These lines of code filter the salaries DataFrame by removing rows where any of the columns from index 7 onwards contain either 0 or -618.13.

salaries1 = salaries.drop(['Notes', 'OvertimePay', 'OtherPay', 'Benefits', 'BasePay', 'TotalPayBenefits', 'Status', 'EmployeeName', 'Agency', 'Year', 'Id'], axis=1)

This code creates a new DataFrame named salaries1 by dropping specific columns from the salaries DataFrame. It removes columns such as 'Notes', 'OvertimePay', 'OtherPay', 'Benefits', and more.

pd.isnull(salaries1).sum()

This line checks for missing (NaN) values in the salaries1 DataFrame and calculates the sum of missing values for each column.

salaries1.dropna(inplace=True)

Here, missing values in the salaries1 DataFrame are removed by using the dropna method with inplace=True, which means the changes are applied to the DataFrame itself.

inputs = salaries1.drop(['TotalPay'], axis=1)
target = salaries1['TotalPay']

These lines split the data into input features (inputs) and the target variable (target). The input features include all columns except 'TotalPay,' and the target variable is 'TotalPay' itself.

from sklearn.preprocessing import LabelEncoder

le_JobTitle = LabelEncoder()

inputs['JobTitle_new'] = le_JobTitle.fit_transform(inputs['JobTitle'])

This code snippet imports LabelEncoder from scikit-learn and uses it to encode the 'JobTitle' column in the inputs DataFrame. The encoded values are stored in a new column called 'JobTitle_new.'

inputs_new = inputs.drop(['JobTitle'], axis=1)

This line removes the original 'JobTitle' column from the inputs DataFrame, leaving only the encoded 'JobTitle_new' column.

from sklearn import tree
model = tree.DecisionTreeRegressor()
model.fit(inputs_new, target)

These lines import the DecisionTreeRegressor from scikit-learn and create a Decision Tree Regression model. The model is trained on the inputs_new data and the target variable (total pay).

model.score(inputs_new, target)

This code calculates the R-squared score of the trained Decision Tree model, indicating how well the model fits the data.

model.predict([[3]])
model.predict([[293]])

These lines make predictions using the trained Decision Tree model. They predict the total pay for employees with certain input features. The first line predicts the total pay for an employee with an encoded 'JobTitle_new' value of 3, and the second line predicts it for a value of 293.

The code essentially preprocesses salary data, encodes categorical features, trains a Decision Tree Regression model, and makes predictions based on the model. It's a basic example of using machine learning for salary prediction.