Update README.md

Added the source of my data set

Author: DerikVo

README.md

@@ -48,7 +48,7 @@ For the purposes of this model, we want to limit our false negatives. We want to
 
 ## EDA
 
-For our [EDA process](./Notebooks/01_EDA.ipynb) we examined our two datasets which contained the classes 'glioma', 'meningioma', 'notumor', 'pituitary' which were separated into their own folders. The file count of the images were as follows:
+For our [EDA process](./Notebooks/01_EDA.ipynb) we examined our two datasets collected from [Kaggle](https://www.kaggle.com/datasets/masoudnickparvar/brain-tumor-mri-dataset) which contained the classes 'glioma', 'meningioma', 'notumor', 'pituitary' which were separated into their own folders. The file count of the images were as follows:
 
 |Class|Training count| Testing count|
 |------|------|------|
@@ -98,4 +98,4 @@ In the future the streamlit app can be use to dynamically evaluate models by all
 
 ## Limitations
 
-[Morgan (2022)](https://www.cancer.gov/rare-brain-spine-tumor/blog/2022/neuroradiology) brings up the argument of variation in MRI scans depending on the machine used. Because we don't know what machine was used to get these MRI scans, we cant account for variations in these images. The most significant difference is the resolution of images between machines, which will affect how well our model learns details. This fact is expended on by [Zacharki et al. (2009)](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2863141/) which explains brain scans are often difficult to distinguish between the tissues of the brain. So in essence the quality of the machine has varying levels of image clarity which would influence our models performance. One area to EDA area to explore is to see the average pixel density of the images or identify what machine was used for the MRI scans.
+[Morgan (2022)](https://www.cancer.gov/rare-brain-spine-tumor/blog/2022/neuroradiology) brings up the argument of variation in MRI scans depending on the machine used. Because we don't know what machine was used to get these MRI scans, we cant account for variations in these images. The most significant difference is the resolution of images between machines, which will affect how well our model learns details. This fact is expended on by [Zacharki et al. (2009)](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2863141/) which explains brain scans are often difficult to distinguish between the tissues of the brain. So in essence the quality of the machine has varying levels of image clarity which would influence our models performance. One area to EDA area to explore is to see the average pixel density of the images or identify what machine was used for the MRI scans.