Hassan Islam

As a master's graduate in Data Science from Kingston University, I am keen to bring over two years of experience as a Junior Data Scientist and analytics expertise in a leading organization. My diverse internship background has sharpened my skills in machine learning, statistical analysis, and data visualization, enabling me to drive insights and inform strategic decisions. I specialize in extracting actionable insights from complex data, and my dedication to problem-solving prepares me to impact a dynamic team and contribute to business growth.

Sentiment analysis of financial News Headlines (Master Dissertation):

• Utilized an extensive dataset from Reuters News headlines for sentiment analysis of financial news.
• Employed SpaCy for named entity recognition and fine-tuned the RoBERTa model to achieve a sentiment analysis prediction accuracy of approximately 97.8%.
• Extended the analysis to predict stock prices, comparing predictions with and without daily sentiment scores.
• Utilized LSTM and GRU models to evaluate the impact of news sentiment scores on stock price prediction, achieving an average accuracy of 97% across all evaluation metrics.

Overall brain survival prediction from MRI in Glioblastoma (BSc Final year project):

• Utilized the BraTS(2018) dataset to segment different types of brain tumours, including Edema, enhancing tumour core, and Necrotic tumour core, using 3D U-NET architecture from MRI images.
• Employed Bias Field Correction techniques for preprocessing MRI images to enhance segmentation accuracy.
• Implemented 3D UNet architecture to accurately segment brain tumour regions and determine their sizes.
• Predicted patients' overall survival rates directly from MRI input images, achieving an overall prediction accuracy of 83%.