CE59700_HW5

Authors:

Pin-Ching Li

Owners:

Pin-Ching Li

Type:

Resource

Storage:

The size of this resource is 435.5 KB

Created:

Apr 28, 2020 at 12:07 a.m.

Last updated:

Apr 28, 2020 at 12:24 a.m. (Metadata update)

Published date:

Apr 28, 2020 at 12:24 a.m.

DOI:

10.4211/hs.203aa39e82b9423684e754a001530b71

Citation:

See how to cite this resource

Sharing Status:	Published
Views:	1622
Downloads:	29
+1 Votes:	Be the first one to this.
Comments:	2 comments

Abstract

This Resource is created for Homework 5 of CE 59700, course held at 2020 Spring semester in Purdue University. The goal of this homework is to forecast streamflow series by a first order exponential model. USGS gage 03335500, Wabash River at Lafayette, is the target gagee in this homework. Our streamflow forecasting model is built by object-oriented programming skill. The error of prediction is reported with respect to different parameter of model. The further optimization process would be taken in the following class of CE 59700.

Task of HW5 is to split your streamflow dataset into training set and validation set. The prediction made by your model is compared to the validation set split in the beginning. This procedure is known as the training process, which is popular in optimization problem. Create a function to report the mean square error of your model. Fit your model with training set and change alpha value manually. Find the best model which returns the minimum MSE value. Turn in the alpha value and the MSE value you get.