This project addresses the issue of inappropriate approvals or rejections by IT/Software Admin Approvers within a corporation, potentially leading to unauthorized access to sensitive resources. It aims to automate the approval process and optimize resource allocation based on historical approval data. The project was segmented into two main parts:

• Mass Approval Detection: An automatic system that classifies approval requests. The process began with data cleaning and balancing the classes in the dataset, followed by the implementation of various classification models. The most impactful features were identified and used to enhance model performance, culminating in the adoption of a voting classifier.

• Resource Allocation: Development of recommendation systems using neural networks, statistical methods, and a random forest classifier. The final model utilizes a random forest to generate recommendations based on encoded application data, predicting approval probabilities which inform resource allocation decisions.