- Conducted data cleaning and data preprocessing on the network traffic dataset from Wireshark
- Performed Exploratory Data Analysis to gain insights on relationship between the independent features and the target value
- Trained and evaluated models using Logistic Regression, XGBost and Dense Netwoork to predict DDoS Attacks
- Optimized models with applying various hyperparameter tuning techniques
Data Science | Hotel Review Sentiment Analysis
- Analyzed hotel reviews to identify factors contributing to customer satisfaction and dissatisfaction, aiming to improve hotel services and ratings
- Utilized Natural Language Processing techniques, specifically Bag of Words, to convert text data into numerical features for sentiment analysis
- Utilized n-grams and stemming for more meaningful multi-word indicators.
- Develop classification models that achieved 78.8% accuracy and an AUC score of 86.7%, indicating good discrimination between positive and negative reviews
- Identified top 20 positive words that reliably indicate guest satisfaction and top 20 negative words that indicate gust complaints
