---
license: mit
language:
  - en
tags:
  - sklearn
  - tensorflow
  - keras
  - random-forest
  - cnn
  - clustering
  - nlp
  - computer-vision
  - recommendation-system
  - time-series
  - streamlit
pipeline_tag: tabular-classification
---

# DataScientst -- 30 Projects, 34 Trained Models

Trained models from a comprehensive ML/AI portfolio spanning 10 categories. All models were trained on real Kaggle datasets.

## Repository Structure

```
regression/          -- Gold, Student, Uber prediction models
classification/      -- Mobile, Wine, Churn classification models
clustering/          -- NBA, Credit Card, Spotify clustering models + scalers
computer_vision/     -- Face mask detection model
nlp/                 -- Spam, IMDb, Fake News models + TF-IDF vectorizers
recommendation/      -- Movie, Book, Music recommendation data + similarity matrices
time_series/         -- Stock, Weather, Store prediction models
data_viz/            -- Visualization CSV datasets
deep_learning/       -- CNN models (Keras) + Markov text generator
metrics/             -- Performance metrics for all models (JSON, CSV, PNG)
```

## Models & Metrics

### Regression
| Model | File | Metric |
|-------|------|--------|
| Gold Price Prediction | `regression/gold_model.pkl` | R² = 0.990 |
| Student Exam Score | `regression/student_model.pkl` | R² = 0.849 |
| Uber/Taxi Fare | `regression/uber_model.pkl` | R² = 0.778 |

### Classification
| Model | File | Metric |
|-------|------|--------|
| Mobile Price Segment | `classification/mobile_model.pkl` | Accuracy = 81.2% |
| Wine Quality | `classification/wine_model.pkl` | Accuracy = 67.5% |
| Customer Churn | `classification/churn_model.pkl` | Accuracy = 78.9% |

### Clustering
| Model | File | Metric |
|-------|------|--------|
| NBA Player Clustering | `clustering/nba_model.pkl` | Silhouette = 0.452 |
| Credit Card Segmentation | `clustering/cc_model.pkl` | Silhouette = 0.531 |
| Spotify Song Clustering | `clustering/spotify_model.pkl` | Silhouette = 0.327 |

### NLP
| Model | File | Metric |
|-------|------|--------|
| SMS Spam Detection | `nlp/spam_model.pkl` | Accuracy = 98.0% |
| IMDb Sentiment Analysis | `nlp/imdb_model.pkl` | Accuracy = 87.3% |
| Fake News Detection | `nlp/news_model.pkl` | Accuracy = 97.6% |

### Deep Learning
| Model | File | Metric |
|-------|------|--------|
| Pneumonia Detection (CNN) | `deep_learning/pneumonia_model.keras` | Val Acc = 92.5% |
| Facial Emotion Recognition (CNN) | `deep_learning/fer_model.keras` | Val Acc = 65.4% |

## Usage

```python
from huggingface_hub import hf_hub_download
import joblib

# Download from a category folder
model_path = hf_hub_download(
    repo_id="OKTAYBBS/DataScientst-models",
    filename="regression/gold_model.pkl"
)

model = joblib.load(model_path)
prediction = model.predict([[1500, 70, 20, 1.1]])
```

```python
# For Keras models
import tensorflow as tf

model_path = hf_hub_download(
    repo_id="OKTAYBBS/DataScientst-models",
    filename="deep_learning/pneumonia_model.keras"
)
model = tf.keras.models.load_model(model_path)
```

## Links

- **Live Demo:** [Streamlit App](https://oktaybobus-datascientst.streamlit.app)
- **Portfolio:** [HF Space](https://huggingface.co/spaces/OKTAYBBS/DataScientst)
- **Source Code:** [GitHub](https://github.com/oktaybobus/DataScientst)