Elise Tshinyangu Mpatu

Driven graduate student leveraging studies in computer science seeks real-world experience as software developer or data scientist. Offers strong interpersonal and task prioritization skills.

Data Mining Projects

• Sentiment Analysis of University of Pretoria using Logistic Regression (Python)

The main objective of this project was to collect articles that talk about the University of Pretoria onnews websites and build a model that will classify those articles into positive or negative categoriesbased on the sentiment expressed in the article. Built a logistic regression model with 89.76% of accuracy using Python. The model was able to predict the sentiment in a article and classify it. A diagram was plotted showing which media had provided more articles about the university and which one had the mostpositive or negative articles.

• Prediction on individual income using Decision Tree Classifier (Python)

The aim of this project was to provide a solution to the problem of income equality. A classification was made to predict whether a person’s annual income is included in the income category greater than $ 50,000 or less than $ 50,000 based on a particular set of attributes. A Decision tree Classifier model was implemented, with an accuracy of 100% on the training data set and 81% on the testing dataset.

Software Engineering Project

• Doctor Appointment system

This project was to design a model of a Health software system - this model is mimicking various aspect of the system. The system uses more than one component which operate in relationship. The aim of the system is to manage the information about a patient attending a heath clinics. People can then go to the website of the health system to register as a patient or a doctor and make an online appointment.

Computer Security Projects

• Anonymized message broadcast system

Design, Develop and Implement. The aim of this project was to create an anonymous email address and send email to one or many person. The recipients will only be able to read the message but cannot reply to the email address , when they try to reply, a message “ address not found” will appear on their screen. The system was configured in a way to disable the option of replying to the email address which will make the address email anonymous and untraceable.

• Continuous Authentication System on mobile devices

In this project, we presented an approach for continuous authentication system on mobile devices. The idea is to implement a solution that will create a database template of the ten user’s fingerprints. The system will then monitor in a continuous way the user touch-entries on the whole screen of the device while he is using it. Those inputs will then be captured, computed with Neural Network algorithm classifier and compared to existing ones to ensure if it is still the legitimate user behind the screen. One of the limitations of this solution is that when it comes to users who use gloves in cold weather, the system will be unable to capture an image of a fingerprint. To counter that it will be necessary to combine fingerprint and another biometric trait.

• Member of research group Data Science for Social Impact by Dr. Vukosi Marivate https://dsfsi.github.io . Department of Computer Science at the University of Pretoria. (2021)
• Treasurer of Congolese Students Society (CSS), University of Pretoria, South Africa (2020)
• General secretary of Women with a Difference Movement, University Protestante of Lubumbashi (2015)
• €“ Senior Female Perfect at College saint Guillaume (2010)