Hugging Face's logo

Mageswaran
/
rf_48_sectors

English
random-forest
classification
bert
sector-classification
machine-learning
Model card Files Community
rf_48_sectors / README.md
Mageswaran's picture
Mageswaran
Update README.md
a5170ac verified
preview code
|
raw
history blame contribute delete
5.54 kB
metadata
language: en
license: other
tags:
  - random-forest
  - classification
  - bert
  - sector-classification
  - machine-learning
inference: false
datasets:
  - custom
model-index:
  - name: RF 48 Sectors Classification Model
    results: []

RF 48 Sectors Classification Model

Overview

This machine learning model is a Random Forest classifier designed to categorize datasets into 48 predefined sectors based on column names. By leveraging BERT embeddings and a sophisticated Random Forest algorithm, the model provides intelligent sector classification for various types of datasets.

Model Details

  • Model Type: Random Forest Classifier
  • Embedding Method: BERT (bert-base-uncased)
  • Number of Sectors: 48
  • Classification Approach: Column name embedding and prediction

48 Supported Sectors

The model can classify datasets into the following sectors:

  1. Agriculture Sector

    • Crop Production
    • Livestock Farming
    • Agricultural Equipment
    • Agri-tech

  2. Banking & Finance Sector

    • Retail Banking
    • Corporate Banking
    • Investment Banking
    • Digital Banking
    • Asset Management
    • Securities & Investments
    • Financial Planning & Advice

  3. Construction & Infrastructure

    • Residential Construction
    • Commercial Construction
    • Industrial Construction
    • Infrastructure

  4. Consulting Sector

    • Management Consulting
    • IT Consulting
    • Human Resources Consulting
    • Legal Consulting

  5. Education Sector

    • Early Childhood Education
    • Primary & Secondary Education
    • Higher Education
    • Adult Education & Vocational Training

  6. Engineering Sector

    • Civil Engineering
    • Mechanical Engineering
    • Electrical Engineering
    • Chemical Engineering

  7. Entertainment & Media

    • Film & Television
    • Music Industry
    • Video Games
    • Live Events

  8. Environmental Sector

    • Environmental Protection
    • Waste Management
    • Renewable Energy
    • Wildlife Conservation

  9. Insurance Sector

    • General Insurance Services
    • Life Insurance
    • Health Insurance
    • Property & Casualty Insurance
    • Reinsurance

  10. Food Industry

    • Food Processing
    • Food Retail
    • Food Services
    • Food Safety & Quality Control

  11. Healthcare Sector

    • Hospitals
    • Clinics & Outpatient Care
    • Pharmaceuticals
    • Medical Equipment & Supplies

Installation 

pip install transformers torch joblib scikit-learn

Usage 

from transformers import BertTokenizer, BertModel
import joblib
import torch

# Initialize model
bert_tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
bert_model = BertModel.from_pretrained('bert-base-uncased', ignore_mismatched_sizes=True)

# Download and load the Random Forest model
model_path = hf_hub_download(repo_id="Mageswaran/rf_48_sectors", filename="model_48_sectors.pkl")
label_encoder_path = hf_hub_download(repo_id="Mageswaran/rf_48_sectors", filename="label_encoder_48_sectors.pkl")

rf = joblib.load(model_path)
label_encoder = joblib.load(label_encoder_path)

def predict_sector(column_names):
    # Convert column names to BERT embeddings
    embeddings = get_bert_embeddings([column_names])
    
    # Predict sector
    prediction = rf.predict(embeddings)
    return label_encoder.inverse_transform(prediction)[0]

# Example
column_names = "clinical_trail_duration, computer_analysis_score, customer_feedback_score"
sector = predict_sector(column_names)
print(f"Predicted Sector: {sector}")

Model Performance

  • Embedding Technique: BERT embeddings from 'bert-base-uncased'
  • Classification Algorithm: Random Forest
  • Unique Feature: Sector classification based on column name semantics

Limitations

  • Model performance depends on the semantic similarity of column names to training data
  • Works best with column names that clearly represent the dataset's domain
  • Requires careful preprocessing of column names

Contributing

Contributions, issues, and feature requests are welcome! Feel free to check the issues page.

License and Usage Restrictions

Proprietary Usage Policy

IMPORTANT: This model is NOT freely available for unrestricted use.

Usage Restrictions

  • Prior written permission is REQUIRED before using this model
  • Commercial use is strictly prohibited without explicit authorization
  • Academic or research use requires formal permission from the model's creator
  • Unauthorized use, distribution, or reproduction is prohibited

Licensing Terms

  • This model is protected under proprietary intellectual property rights
  • Any use of the model requires a formal licensing agreement
  • Contact the model's creator for licensing inquiries and permissions

Permissions and Inquiries

To request permission for model usage, please contact:

  • Email: [Your Contact Email]
  • Hugging Face Profile: [Your Hugging Face Profile URL]

Unauthorized use will result in legal action.

Contact

meyyappanmageswaran@gmail.com

Citing this Model

If you use this model in your research, please cite it using the following BibTeX entry:

@misc{mageswaran_rf_48_sectors,
  title = {Random Forest 48 Sectors Classification Model},
  author = {Mageswaran},
  year = {2024},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/Mageswaran/rf_48_sectors}}
}

Additional Resources

Acknowledgments

  • Hugging Face Transformers