Employee Salary Prediction Model [Supervised Machine Learning method]

This repository contains Employee salary prediction machine learning model developed using Supervised machine learning method. A stramlit app is also tailored to the model for better user interface.

The model is developed in google colab

The Objective for building his model

• The objective is to develop an Employee Salary Prediction Model and evaluate various supervised machine learning models to accurately predict employee salaries based on a comprehensive set of relevant attributes. The aim is to identify the most effective algorithm that yields the highest predictive accuracy and provides actionable insights into the factors influencing employee compensation within the organization.

Libraries used:

• pandas – for Data manipulation and analysis
• matplotlib – for Data visualization, plotting graphs
• sklearn – for Machine learning toolkit
• sklearn.preproceesing – for data preparation, scaling, encoding
• sklearn.pipeline – for workflow automation, model chaining
• sklearn.model_selection – for model evaluation, splitting data
• sklearn.linear_model – for linear regression model
• sklearn.ensemble – for Ensemble methods
• sklearn.neighbors – for K-Nearest Neighbors
• sklearn.svm – for support vector machines
• joblib – for Model saving and loading
• stramlit – for web app creation for ML

Algorithms used:

• LogisticRegression
• RandomForest
• KNN
• SVM
• GdientBoosting

Procedure (steps involved):

• Phase 1: Data Acquisition and Initial Assessment
• Phase 2: Data Preprocessing and refinement
• Phase 3: Model Development and Evaluation
• Phase 4: Application Development and Deplyoment

Results:

Models Prediction values (graph):

Model's APP intrface (created with Streamlit):

Note:

1. The jupitor file can be downloaded and directly deployed in the Google colab or any susequent platforms for testing.
2. For more information about procedure or results, please do check the presentation.
3. The authtoken for the streamlit app is taken from ngrok [if you are developing model in your local machine's jupitar notebook then you need not to worry about the ngrok authtoken, you can skip this part].