Muskaan Patel's Projects
Basic Machine Learning Projects based on Regression, Classification, Clustering etc. along with their Databases.
MLND-Capstone My project is a simple Credit Card Approval Prediction using Machine Learning. It implements Binary Classification using LinearLearner in AWS SageMaker environment. My capstone project for Udacity's Machine Learning Nanodegree. The folder includes my initial project proposal, my final project report, the csv file used, and the final project ipynb(jupyter notebook) file. It also includes a directory called CreditCardApproval which contains 4 file. Two of them are rough files and not meant to be used. The other two are test.csv and train.csv which are the final files I upload on S3 after all the pre-processing. No as such special environment/library is required for the implementation of this project.
Density-based spatial clustering of applications with noise is a data clustering algorithm.
Final Project under CS2470 Deep Learning Course (Brown University - Master's in Computer Science)
Detection &Ā Classification of Benign and Malignant Cancer using Image Processing. Additionally, a basic CNN approach for the same.
Practising BERT FineTuning for Question and Answering task.
Attempt to finetune a Mistral 7B model using a Medium tutorial. Link:https://medium.com/@hugo_fernandez/fine-tune-and-deploy-an-llm-on-google-colab-notebook-with-qlora-and-vertexai-58a838a63845
Working on a loan test and train dataset. Applying various algorithms such as K Nearest Neighbour, Decision Tree, Support Vector Machine and Logistic Regression. Various Metrics such as Jaccard Index, F1 Score and Log Loss is also used. All the algorithms are tested and manipulated in order to achieve maximum accuracy.
Config files for my GitHub profile.
Firstly, the two libraries, i.e Numpy and OpenCV are imported.Then the video is captured. If there is a destination path assigned to the VideoCapture function then that video is captured, else if 0 is mentioned then live input is taken from the webcam itself. Next, the createBackgroundSubtractorMOG2 function is used to create a mixture-of-gaussian background subtractor. This function adapts itself to subtract the background with the last frames. This function also automatically applies Gaussian filtering and some morphological filtering to reduce noise and can also detect shadows, The history parameter indicates the length of the history. A while loop is then introduced in which cap.read() is executed which returns a bool, if frame is read correctly. Once we have the frames we can use the subtractor to find get the background. And lastly we display everything.
PoseNet - Final Project under CS1430 Computer Vision Course (Brown University - Master's in Computer Science)
Practising PyTorch/TensorFlow Pipelines in context to DL, ML, etc.
It uses Image Processing and Deep Learning Techniques to recognize any anomalous activity in a live CCTV footage, thereby sending an alert to the security team.