Pratiksha Pradhan

● Designed conceptual data models and mapped it to a relational model to build the database for a recommerce platform ● Executed relational model via MySQL and non-relational model via NoSQL in MongoDB to query data ● Accessed the database via Python and created Matplotlib visualizations to gain insights such as monthly change in new clients

● Conducted data preprocessing, used librosa and ffmpeg to produce functions to extract features from each audio sample ● Developed a deep-feed forward neural network model with five hidden layers to predict the gender of the voice input ● Facilitated the testing of the model by inputting own voice using torchaudio, IPyWebRTC, IPython, with an accuracy of 85%

• Performed data visualization, oversampling, data partitioning, standardization and scaling, and dimension reduction on dataset • Implemented machine learning models like logistic regression, gradient boost, classification trees, Naïve Bayes using sklearn along with hyperparameter tuning, recursive feature elimination, k-fold cross validation to find the best classification model • Evaluated and visualized model performance to conclude that gradient boost is the best algorithm with a sensitivity of 90%