CBSA-Data-Analytics-Challenge

Link: https://docs.google.com/document/d/1RiR0BsdOug2UZ822Up0-FkCEGcv8ItKsyChahozDrOo/edit

---

Workflow:

1. Data Exploration

   • {wordcloud}

2. Preprocess data

   • Tokenization
   • Remove useless data (manual)
   • Remove stop words
   • Remove symbols

3. Feature Extraction / Word Embeddings

   • Tf-idf
   • Sentence Transformer

4. Classification

   • Naive bayes
   • SVM
   • CNN (Optional)
   • LSTM (Optional)

5. Regression

   • CNN
   • LSTM

6. Unsupervised

   • {lda}

7. Data Visualization

   • {2d dimension}
   • {model comparison}

8. Others

   • Imbalanced Data

---

Deck Preparation:

1. Intro
2. Pain points, Why this prediction task, context
3. Importance
4. Data Exploration
5. Methodology:
   • tried what model
6. Model results
   • accuracy
   • training time
   • space complexity (Optional)
7. Limitations and Improvements
8. Prototype (Streamlit) (Optional)

---