Bi-Weekly Report # 1

On February 23, 2023, our team convened to discuss our research direction in Behavioral Economics. During the meeting, we reviewed a paper titled "Machiavelli Preferences Without Blame: Delegating Selfish vs. Generous Decisions in Dictator Games" authored by Glynis Gawn and Robert Innes. We also searched for datasets related to this topic, but discovered that the existing data was challenging to analyze predictively.

Subsequently, we shifted our focus to exploring other psychology-related datasets, specifically those related to lying behavior and deception. Finally, we turned our attention to brain imaging, and became intrigued by the potential for studying brain tumors and MRIs in general. After conducting a thorough search, we were able to identify a promising dataset on Kaggle.

The dataset is owned by Fernando Feltrin and contains MRI (Magnetic Resonance Imaging) images of brain tumors. The dataset can be found here. The dataset has a total of 4479 images, each of which is a 2D slice of a patient's brain, and it has been collected from real patients diagnosed with brain tumors. The dataset is divided into 14 classes of brain tumors: astrocytoma, carcinoma, ependymoma, ganglioglioma, germinoma, glioblastoma, granuloma, medulloblastoma, meningioma, neurocytoma, oligodendroglioma, papilloma, schwannoma and tuberculoma. The images in this dataset have been preprocessed and cropped to remove irrelevant parts of the MRI scans. They are stored in JPEG format. Additionally, the images are labeled with the corresponding tumor type.

We found another dataset on brain tumors using MRI scans. The same is linked here. The dataset is divided into two categories based on the presence or absence of brain tumors. The "yes" category contains 155 MRI images with brain tumors, and the "no" category contains 98 MRI images without brain tumors.

These datasets are useful for us because we are interested in developing algorithms to accurately detect and classify different types of brain tumors from MRI scans. Either dataset can be used to train and validate machine learning models (like PCR, QDA, or LDA) or deep learning models, which could help improve the accuracy and speed of brain tumor diagnosis. 

Suryadyuti: Lead Researcher

Simeon: Lead Programmer

Jason: Assistant Programmer and Researcher

Kenza: Data Aggregation and Organization

Tanay: Model Validation and Visualization