ABSTRACT:

Voting is an important Ensemble Learning technique. However, there has not been much discussion about leveraging the base classifiers’ consensus on unlabeled data to better inform the final prediction. My proposed method identifies the data points where the ensemble reaches consensus and where conflict arises in the unlabeled space. A meta weighted KNN model is trained upon this half-labeled set with the labels of the consensus and the conflict points marked as “Unknown,” which is treated as a new, additional class. The predictions of the meta model are expected to better inform the decision of the ensemble in the case of conflict. This research project aims to implement my proposed method and evaluate it on a range of benchmark datasets.

SOFTWARE ARCHITECTURE DIAGRAM:

FINAL PAPER

SOFTWARE DEMO VIDEO

Chest X-Ray Abnormalities Detection with a focus on Infiltration

Abstract

Navigating chest X-rays is an obligatory step to determine lung and heart diseases. Since many people now believe that Chest X-ray radiographs can detect COVID-19, the disease of the decade, the problem of Chest X-ray Abnormalities Detection has gained increasing attention from researchers. Numerous machine learning algorithms have been developed to address this problem to raise reading accuracy, improve efficiency, and save time for both doctors and patients. In this work, I propose a model to determine whether a Chest X-ray image has Infiltration and to detect the abnormalities in that image using YOLOv3. The model will be trained and tested with the VinBigData dataset. Overall, I will use the existing tool, YOLOv3, on a new problem of detecting Infiltration in Chest X-ray radiographs.

Paper

Software Architecture Diagram

Project abstract:

Heuristic evaluation is one of the most popular methods for usability evaluation in Human-Computer Interaction (HCI). However, most heuristic models only provide generic heuristic evaluation for the entire application as a whole, even though different parts of an application might serve users in different contexts, leading HCI practitioners to miss out on context-dependent usability issues. As a prime example, mobile search interfaces in e-learning applications have not received a lot of attention when it comes to usability evaluation. This paper addresses this problem by proposing a more domain-specific and context-dependent heuristic evaluation model for search interfaces in mobile e-learning applications. I will analyze studies on mobile evaluation heuristics, in combination with research in mobile search computing and e-learning heuristics, to generate a heuristic model for mobile e-learning search interfaces.

Software architecture diagram

Research paper

https://drive.google.com/file/d/17a4p9N_OzE-O7VrwpSl7a9cbAI4Lx_25/view?usp=sharing

Software demonstration video

Predicting cryptocurrency price movements is a well-known problem of interest. In this modern age, social media represents the public sentiment about current events. Twitter especially has attracted a lot of attention from researchers who are studying the public sentiments. Recent studies in natural language processing develop automatic techniques in analyzing sentiment in social media information. This research is directed towards predicting volatile price movement of cryptocurrency by analyzing the sentiment on social media and finding the correlation between them. Machine learning algorithms including support vector machine and linear regression will be used to predict the prices. The most efficient combination of machine learning algorithms and the datasets being used will be determined.

Software Architecture Diagram

Link to video tutorial: https://youtu.be/ml4Tc-Xr7bc

Link to senior paper: https://drive.google.com/file/d/1S2TUrBGu8VMmVX1iES8osTWCS6J3PfOG/view?usp=sharing

• Software demo video
• Software Architecture Diagram:

dongcao_research_paper_draft1.pdf