Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/dearabhin/girlfriend-predictor

Using machine learning to solve the ultimate college classification problem. A fun project applying Python and Logistic Regression to predict relationship outcomes based on a (hilariously) synthetic dataset. 📊❤️
https://github.com/dearabhin/girlfriend-predictor

classification data-science fun-project google-colab jyputer-notebook jypyternotebook logistic-regression machine-learning pandas python scikit-learn

Last synced: about 19 hours ago
JSON representation

Using machine learning to solve the ultimate college classification problem. A fun project applying Python and Logistic Regression to predict relationship outcomes based on a (hilariously) synthetic dataset. 📊❤️

Awesome Lists containing this project

README 

          
# Girlfriend Predictor V3: An Advanced Machine Learning Project



*An engineering student's fun take on solving life's most pressing classification problem, now updated with advanced social, academic, and environmental features for a more realistic model.*



[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1HCgPxf3gdnqQyZgHvqao9dZHn53QUv7D?usp=sharing)



---





  Girlfriend Predictor




---



## 🧐 Overview



As a B.Tech student at Model Engineering College, I decided to apply machine learning to a critical classification problem: **Will I get a girlfriend in college?**



This is **Version 3** of the project, which moves beyond simple metrics to include a more nuanced and holistic set of features. It now considers academic performance, collaborative spirit, direct communication efforts, and even environmental factors like the "local competition." The result is a more robust and insightful model that better reflects the complexities of real-life social dynamics, all while serving as a practical application of data science principles.



## ✨ Advanced Features (V3)



The V3 model is trained on a richer, more sophisticated synthetic dataset. The features are designed to be more realistic and quantifiable aspects of a student's life.



#### Personal & Habitual Features

* `personal_grooming_score`: (Scale 1-10) An upgrade from the original `uniform_wrinkle_score`. A higher score indicates better personal presentation and outlook.

* `communication_skill_score`: (Scale 1-10) A self-assessed score for soft skills, directly modeling the importance of effective communication.

* `avg_dm_conversations_per_week`: (Number) Quantifies direct digital communication efforts across platforms like Instagram and WhatsApp.

* `crush_eye_contact_count`: (Number) The classic, powerful, but often sparse feature indicating mutual interest.



#### Academic & Technical Features

* `semester_cgpa`: (Number) Your current SGPA/CGPA, serving as a direct measure of academic performance.

* `hackathon_participation_count`: (Number) An upgrade from "wins" to "participation," this shows initiative and passion for technology.

* `other_competition_wins`: (Number) Captures success in other competitive events like coding contests, paper presentations, etc.

* `provides_academic_help`: (Binary: 1=Yes, 0=No) A measure of collaborative spirit—do you help others with assignments, doubts, or projects?



#### Social & Environmental Features

* `events_attended`: (Number) The number of workshops, tech fests, and cultural events you show up to.

* `dept_similarity_index`: (Binary: 1=Yes, 0=No) A simple flag to check if you're in the same department.

* `friend_zone_flag`: (Binary: 1=Yes, 0=No) The classic categorical label with a heavy negative penalty in the model.

* `local_competition_score`: (Scale 1-10) An environmental factor measuring how many "competitors" are also vying for the same person's attention.

* **`target_is_single`**: (Binary: 1=Yes, 0=No) The **game-changer**. This feature introduces a hard, real-world constraint. If the target person is not single, the model learns that a successful outcome is nearly impossible.



## 🔬 Methodology & Concepts



This project follows a standard machine learning pipeline.



#### 1. Logistic Regression

This project uses **Logistic Regression**, a fundamental algorithm for binary classification. It's used to predict a probability (a value between 0 and 1) for a given class, making it perfect for "Yes/No" questions like ours.



#### 2. Feature Scaling (`StandardScaler`)

With many new features on different scales (e.g., `CGPA` from 0-10 vs `events_attended`), scaling is crucial. **Standard Scaling** transforms the data to have a mean of 0 and a standard deviation of 1, ensuring all features contribute fairly to the model's predictions.



#### 3. Model Evaluation

We assess the model's performance using standard metrics:

* **Accuracy**: The percentage of correct predictions.

* **Classification Report**: Shows Precision, Recall, and F1-Score.

* **Confusion Matrix**: Visualizes the model's correct and incorrect predictions (True Positives, False Negatives, etc.).



## 🚀 How to Run



1.  **Clone the repository:**

    ```bash

    git clone https://github.com/dearabhin/girlfriend-predictor.git

    cd girlfriend-predictor

    ```

2.  **Set up a virtual environment and install dependencies:**

    ```bash

    pip install -r requirements.txt

    ```

3.  **Run the script or notebook:**

    * Open and run the `girlfriend_predictor.ipynb` in Google Colab or Jupyter.

    * Alternatively, run the Python script from your terminal: `python girlfriend_predictor.py`



## 🤖 The V3 Simulator

In the code, you can find the `your_current_stats_v3` dictionary. Change the values to reflect your own situation and see what the advanced model predicts for you! Experiment with changing `target_is_single` from 1 to 0 to see its powerful effect.



```python

# --- YOUR TURN: Fill in your stats for the V3 model! ---

your_current_stats_v3 = {

    'events_attended': 15,

    'semester_cgpa': 9.1,

    'provides_academic_help': 1,

    'target_is_single': 1,  # Try changing this to 0!

    'local_competition_score': 6,

    # ... and all the other new features

}