Made by:

---

- Habibun Nabi Hemel

- Ishrak Hamim Mahi

1. Overview:

FutureFit is a predictive modeling tool designed to help BSc students understand their placement prospects and expected salary packages by leveraging machine learning algorithms. The system analyzes historical placement data of students, including factors such as academic performance, skills, internships, and more, to predict whether a student will secure a great, decent, or poor job placement in terms of salary.

The tool's goal is to motivate students to focus on enhancing the skills that will improve their chances of securing higher-paying jobs and to guide them in their preparation for campus placements.

2. Data & Preprocessing:

The dataset used for training and testing the machine learning models is synthetic, as no large-scale real-world dataset was available for this type of prediction. The synthetic dataset consists of 220,000 rows and 19 features, representing various attributes such as:

• Academic performance (GPA, subjects studied)
• Internships completed
• Skills acquired (programming languages, certifications)
• Previous placements and salary
• Location and company size

The Dataset's link is https://www.kaggle.com/datasets/firozchowdury/placement-package-ctc-prediction-dataset

Figure 1: DataSet

3. Steps Taken in Preprocessing:

1. Data Cleaning:
   Irrelevant or redundant details were removed from the dataset.

2. Handling Missing Values:
   Missing data points were either filled with the mean or median (for continuous variables) or the most frequent value (for categorical variables).

3. Label Encoding:
   Categorical features (like department, gender, etc.) were converted into numerical format using label encoding, making them suitable for machine learning models.

4. Splitting Data:
   The data was split into training and testing sets (70% for training, 30% for testing) to ensure the models' effectiveness and avoid overfitting.

4. Model Selection & Algorithms:

The system uses several machine learning algorithms to predict placement success and salary brackets. These include:

1. Decision Tree Classifier:
   A tree-like model used to make predictions based on decision rules. It was tested for its ability to capture non-linear relationships between the features.

2. Random Forest Classifier:
   An ensemble of decision trees that helps reduce overfitting by averaging multiple decision trees, making it a more stable and accurate model.

3. Support Vector Machine (SVM):
   A classifier that works well for high-dimensional spaces. It attempts to find the optimal hyperplane that separates different classes.

4. Logistic Regression:
   A simple yet powerful algorithm used for binary or multi-class classification tasks, predicting the probability of placement success.

5. Naïve Bayes:
   A probabilistic classifier based on Bayes' theorem. It’s particularly useful when the features are independent.

6. K-Nearest Neighbors (KNN):
   A non-parametric method that classifies a data point based on how its neighbors are classified.

5. Prediction Classes (Salary Ranges):

The placement outcome is categorized into three salary brackets:

• Great (Above 20 LPA):
  Students who fall in this category are predicted to secure jobs with a salary higher than 20 Lakhs per annum (LPA). These placements are typically with top-tier companies or highly competitive industries.

• Decent (10-20 LPA):
  Students in this category are expected to land jobs with salaries between 10 LPA and 20 LPA. These are considered solid job offers but may not be from the highest-paying companies.

• Poor (Below 10 LPA):
  Students whose placement predictions fall into this category are expected to secure jobs with salaries under 10 LPA, often in smaller or less competitive companies.

6. Model Performance:

After training and evaluating various models, it was found that Random Forest provided the most accurate predictions. This suggests that a combination of decision trees is more capable of understanding the complex relationships between the student’s characteristics and their placement prospects.

7. Live Application:

The FutureFit platform is live online, where students can input their data (such as academic performance, skills, etc.) and receive predictions about their potential placement salary. The platform provides the following:

• Real-time predictions: By simply entering data, students can get immediate feedback about their placement prospects.
• Career Guidance: Students can better understand their strengths and weaknesses and focus on improving specific skills for better placement opportunities.

is live on https://futurefit.streamlit.app/

Webpage

Figure 2: FutureFit is live

Conclusion:

FutureFit serves as a predictive and motivational tool for students, enabling them to make informed decisions regarding their preparation for campus placements. By forecasting potential job salaries, it helps students:

1. Set realistic career goals.
2. Identify key areas for improvement.
3. Gain insights into their future career trajectories.

This system not only encourages students to develop in-demand skills but also empowers them to optimize their preparation, ultimately leading to higher chances of securing better job offers.

Paper

The overleaf link: https://www.overleaf.com/read/bcxpdkysbgbw#86b8d1

Figure 3: paper