This project is interested in solving the Kaggle BCI Challenge @ NER 2015. It showcases efforts in cleaning the noisy data and training multiple models that classify whether a feedback was good or bad. The data consists of 56 P300 electrode signals across the brain, EEG evoked responses recorded every 5 milliseconds. The goal of the competition is to detect errors during the spelling task, given the subject’s brain waves.
With E-commerce businesses such as Amazon and eBay seeing an unprecedented rise of online sales, the safety of mailmen and customers residing in their homes must not be overlooked. Consequently, for our project, we chose to extend the iOS/Android Amazon mobile application to implement COVID-19 guidelines. While most package deliveries do not require customers’ signature, some of the more valued items require that the resident signs for the package at the time of delivery. With our UI extension to the app, safer and efficient delivery can be provided and help prevent the unnecessary spread of COVID-19.
This data science project tries to identify whether intrastate travel has any effect on the increase of COVID rates per state. We clean and merge two datasets, one containing the daily COVID rates per state and the other summarizing intrastate travel per state. We intend to focus on daily positive increases as a function of intrastate travel, and will explore the data using both visual and predictive analysis.
Neural networks and deep learning algorithms have been acknowledged for demonstrating their strengths in predicting multi-class outputs given original multi-dimensional inputs. This project attempts to create the optimal convolutional neural network for the CIFAR-10 dataset using the neural networks library Keras. A varying selection of filters, activation functions, pooling functions, and optimizers were tried for the learning algorithm. The functionality of each learning algorithm was determined using their corresponding test accuracies, and the algorithm that produced the highest test accuracy is considered to be the optimal convolutional neural network.
Supervised machine learning algorithms have been acknowledged for demonstrating their strengths in predicting outputs given original inputs. This project attempts to reproduce the results found by Caruana and Niculescu-Mizil in their study. The datasets considered in this project pertain to the fields of finance and health science, two of many fields that their success improves our daily life.
I served as the Vice President of Technology Operations in UES for the 2017-2018 school year. There, I revamped the club's website and maintained it per the club's needs.
“Wash your frustrations away with TritonWash. Laundry, reinvented.”
TritonWash will be the only app that a student doing laundry needs at UCSD. With its interactive UI, TritonWash will deliver the most accurate data by displaying which washers/dryers are available to use in every laundry location on campus.