Suicide is a leading cause of death worldwide. This thesis presents the application of Hawkes processes on suicide data from England and Wales since no paper has ever used the Hawkes process model on the analysis of such a topic before. I began by introducing background knowledge and a brief data description. After reviewing the concepts of the Hawkes process and of the general point process, I modified the marked Hawkes process model and applied it, in order to understand the suicide copycat behaviors or even suicide cascade phenomenon based on the data obtained. Two kernel functions were used to compare the predicted results. The Hawkes process model with the power-law kernel seems to provide a relatively more reliable prediction. With predictions, the corresponding preventative interventions were significant as well. Last, I provided an overview of the current analysis and offered further directions and future potential improvements on the explorations of this topic.

This study investigates how individual sentiment towards product help another individual purchasing decision. It explores numerous sentiment analysis model such as Ordinal Regression, LSTM, Logistic Regression, Random Forest, and BERT, in predicting customer recommendations and ratings from Sephora product reviews. These findings demonstrate the models' ability to forecast customer recommendations consistently, offering significant insights that can enhance customer satisfaction, marketing initiatives, and inventory management. Despite the promising results, limitations, such as capturing the nuanced nature of evaluations and ensuring model generalizability, remain, emphasizing opportunities for future research.