My personal data science projects include various topics, including data analyses and prediction models, and applying different machine learning methods such as Support Vector Machine model, Deep Neural Network model, and open source Prophet model.
Each project has raw data files, ipython notebooks, and write-up in PDF format. Projects include:
- Avocado Prices Prediction:
- Data visualization to analyze trend and seasonality factors of avocado prices.
- Facebook's open source Prophet model to predict avocado prices.
- Breast Cancer Diagnosis:
- Transfered learning models such as ResNet18, ResNet34, ResNet50 to classify breast tumors.
- Customized convolutional neural net model.
- Hyper-parameter and loss function tunining.
- Chest X-Ray:
- Customized Convolutional ResNet model.
- Transfered learning ResNet34.
- Diabetes Stage Classification:
- Independent variables include but not limited to gender, age, family income, arm circum.
- Dependent variable indicate diabetes stage (e.g., 0 means no diabetes).
- Multilayer perceptron models.
- House Prices:
- Predicting house prices using features such as kitchen area, year built, garage availability.
- Stacked model including Gradient Boosting, and Lasso lineared regression.
- News Article Popularity:
- Multi-class prediction models to predict popularity of news articles.
- Models include logistic regression, decision tree, xgboost, knn.
- Marketing Brand:
- Data analysis and linear regression model to explain sales in response to marketing plans.