For this problem you use the data in "admData20.csv" (https://drive.google.com/file/d/16lrybf11bpCv7ADlyz9gDNaQgQxweZNm/view?usp=sharing). This file contains information about the admission data for an academic program at a college that has 5 start dates throughout the year. The features in the file include the start date, the weekly statistics on the accumulative number of students who applied and admitted for a specific start date (time-series), the number of weeks before the start date, and the the budget value for the number of admissions for the given start date. The objective of the problem is to predict the accumulative number of ad- missions 1-5 weeks from the current date. i.e. on any given week, you need to forecast what the accumulative number of admissions will be in 1 week, 2 weeks,...5 weeks from that week.

Use any machine learning model (Linear Regression, KNN, Random Forests,

regular deep NN, etc.) or an ensemble of models to achieve the lowest

MAE.

(a) Report the MAE of the test set on your best model.

(b) Plot the loss curves for training and validation sets.