dfavenfre Goto Github PK

I'm a Data Scientist, and currently committing to an MSc. in Applied Data Science. In my repositories, you'd see various projects I personally developed. I also deploy some of my projects on Streamlit.

Here are a few notable projects I have worked on:

• LLMRoboFund: LLMRoboFund is a powerful chatbot empowered with Multi-document RAG. The chatbot is equipped with RetrievalQA and SQL Agents, aiming to ease the investment research.

• Forecasting Hourly Electricity Prices: Developed various predictive models, including a hand-crafted neural network (LSTM) model for time-series forecasting and an XGB regression model to predict t+1 forecast window.

• Econ Dashboard: Created a centralized financial dashboard for screening financial/economic data, sentiment classification, and time-series forecasting. This all-in-one dashboard is equipped with hand-crafted neural network models, including a sentiment classifier with pre-trained embeddings, and LSTM forecast models for different market capitalizations.

• Financial Sentiment Classifier: Developed an NLP model to analyze the sentiment of financial or economic commentary, tweet, or news. It is built upon Universal Sentence Encoder (USE) embedding layer and is fine-tuned for financial sentiment classification purposes. Financial Phrasebank's 'agreeall' dataset was used for fine-tuning.

I occasionally write blog posts on various topics at Medium. Here are some of my recent articles:

• Leveraging Lagged Exogenous Variables For Time-Series Forecasting — Without Time I explore additional usage of ML models to forecast (t+n) horizons with lagged exogenous variables. On top of that, I deploy a neural network time-series model to create a benchmark for later comparison, using nn.LSTM layers.

• House Price Prediction: Stochastic Gradient Boosting With KNN Imputer for Pre-processing: A depth-analysis on Kaggle's House price prediction competition, along with my submission. Within this medium blog, I explore the usage of KNNImputer for missing value imputations, engineer some features, build a model using Stochastic Gradient Boost, conduct hyperparameter tuning deploying Optuna and finaly, I evaluate feature selection based on SHAP algorithm.

• Credit Score Prediction With Multi-Model Ensemble Voting Classifier I share a detailed-out process starting from cleaning and interpolating to building a voting-classifier as an introduction to building a classification model, using various machine learning algorithms.