Assignment Instructions

Assignment%20Instructions

User Manual:

Open the PDF directly: View PDF .
Page Count: 1

Download
Open PDF In Browser	View PDF

Coding Assignment - 1 - Implementation of Multiple linear
Regression using Gradient Descent
Hi Guys,
In this coding assignment you need to use the dataset provided to you in dataset folder and
implement the multiple linear regression model using gradient descent technique.
Problem:
The data you are given is about computer hardware attributes and using those hardware
attributes you are suppose to estimate relative performance of the cpu.
Dataset:
For detailed information on the dataset and attributes you can visit the following link:
https://archive.ics.uci.edu/ml/machine-learning-databases/cpu-performance/machine.names
Specific Instructions:
You need to use following attributes as the input variables (X):
1) MYCT: machine cycle time in nanoseconds
2) MMIN: minimum main memory in kilobytes
3) MMAX: maximum main memory in kilobytes
4) CACH: cache memory in kilobytes
5) CHMIN: minimum channels in units
6) CHMAX: maximum channels in units
And the following attribute as response variable (Y):
1) PRP: published relative performance (integer)
Code:
You need to use the given directory structure and put the code you develop in the code
folder. You need to upload the final version to your github repository for feedback and
evaluation.
Related Articles and documents:
For Gradient descent, you can refer to our sessions/notes as it contains all the necessary
information for you to develop this implementation. Also you can refer to the code of gradient
descent which we used for implementing simple linear regression using gradient descent
Here’s another article you might find helpful:
https://medium.com/deep-math-machine-learning-ai/chapter-1-2-gradient-descent-with-mathd4f2871af402



---

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.5
Linearized                      : Yes
Producer                        : Skia/PDF m69
Page Count                      : 1

EXIF Metadata provided by EXIF.tools