Data Science Portfolio
View the Project on GitHub JakovGotovacBorcic/Internship_Portfolio
#Comparison Of Convolutional Neural Network And Vision Transformer
The goal of this competition was to find a model that would predict emotional expressions from human faces. My team shared the first place with 86% accuracy by computing an array of features and using support vector machines as our predictive model.
Used a Lasso regression model to classify different types of physical activity from signals measured by the accelerometer and gyroscope in smartphones using time and frequency domain features.
Built a model that predicts personality scores of YouTubers based on transcripts of their videos using lexicons for sentiment analysis.
Built a classifier that recognizes positive and negative reviews of Amazon products using Bag-of-Words features.
Eelloo is a company that offers a talent development platform and a recruitment platform for other companies. They help their customers find the best job candidates and they help the companies employees plan their careers. Unfortunately, not all employees use this opportunity so my team’s task was to find the differences between active and inactive users. In less than a week we came up with four different ways to measure user’s activity and tested if user’s personality can predict their activity on the platform. By approaching the problem with multiple possible activity measures we managed to show how important it is to clearly define an active user helping Eelloo get insight into what type of activity is the most important for them.
Language from the domain of space and movement is often used when talking about time. People spontaneously adopt one of two perspectives: an ego-moving perspective in which the subject is approaching a static future or a time-moving perspective in which the future is approaching a static subject. Based on the perspective adopted, it can be said that we are approaching a deadline or that a deadline is approaching. Earlier research has suggested that reporting on uncertain events from a time-moving perspective can cause the event to be perceived as more serious, but also more tractable. The aim of this experimental research is to examine the effect of the time perspective that the participants spontaneously adopt and the time perspective of the article on the perception of the war in Ukraine. Participants (N = 801) filled out a questionnaire via an online link that measured the time perspective they spontaneously adopt, as well as their anxiety and their perception of risk, solvability of the war, inevitability of consequences, and urgency after reading an article about the war written from the time-moving or the ego-moving perspective. Results suggest that the interaction of the spontaneous time perspective and the time perspective of the article affects anxiety and the perception of urgency, while it does not affect the perception of risk, solvability of the war and inevitability of the consequences.