---
language: en
license: mit
tags:
  - tax-compliance
  - financial-compliance
  - machine-learning
  - tax-regulations
model-index:
  - name: Finlytic-Compliance
    results:
      - task:
          type: compliance-check
        dataset:
          name: finlytic-compliance-data
          type: financial-transactions
        metrics:
          - name: Accuracy
            type: accuracy
            value: 86.00
          - name: Precision
            type: precision
            value: 90.00
          - name: Recall
            type: recall
            value: 82.00
          - name: F1-Score
            type: f1
            value: 89.00
        source:
          name: Internal Evaluation
          url: https://huggingface.co/comethrusws/finlytic-compliance
---

# Finlytic-Compliance

**Finlytic-Compliance** is an AI-driven model built to automate the task of ensuring financial transactions meet regulatory tax requirements. It helps SMEs remain compliant with tax laws in Nepal by constantly monitoring financial records.

## Model Details

- **Model Name**: Finlytic-Compliance
- **Model Type**: Compliance Check
- **Framework**: TensorFlow, Scikit-learn, Keras
- **Dataset**: The model is trained on financial transactions labeled for tax compliance.
- **Use Case**: Automating the detection of tax compliance issues for Nepalese SMEs.
- **Hosting**: Huggingface model repository (locally used)

## Objective

The model reduces the need for manual checking and reliance on tax consultants by automatically flagging transactions that do not comply with Nepalese tax laws.

## Model Architecture

The model is built on a transformer architecture, fine-tuned specifically for identifying compliance issues in financial transactions. It has been trained on a dataset of transactions with known compliance statuses.

## How to Use

1. **Installation**: Clone the model repository from Huggingface or load the model locally.
   
   ```bash
   git clone https://huggingface.co/comethrusws/finlytic-compliance
   ```
   
2. **Load the Model**:

   ```python
   from transformers import AutoTokenizer, AutoModel

   tokenizer = AutoTokenizer.from_pretrained("path_to/finlytic-compliance")
   model = AutoModel.from_pretrained("path_to/finlytic-compliance")
   ```

3. **Input**: Feed the model financial transactions (structured in JSON or CSV format). The model will process these transactions and check for compliance issues.

4. **Output**: The output will indicate whether a transaction is compliant with tax regulations and provide additional insights if necessary.

## Dataset

The model was trained using annotated financial records, with transactions labeled as either compliant or non-compliant with Nepalese tax laws.

## Evaluation

The model was evaluated using a hold-out test dataset. The performance metrics are as follows:

- **Accuracy**: 92%
- **Precision**: 90%
- **Recall**: 88%
- **F1-Score**: 89%

These results indicate that the model is highly effective in flagging non-compliant transactions and ensuring financial records are accurate.

## Limitations

- The model is designed for Nepalese tax laws, so it may need adjustments for different regulatory frameworks.
- It is best suited for common financial transactions and may not generalize well for edge cases.

## Future Improvements

- Expanding the dataset to cover more complex financial scenarios.
- Adapting the model to work with tax regulations from other countries.

## Contact

For queries or contributions, reach out to the Finlytic development team at [finlyticdevs@gmail.com](mailto:finlyticdevs@gmail.com).