Download Deep Learning Made Easy With R PDF

TitleDeep Learning Made Easy With R
File Size5.5 MB
Total Pages251
Table of Contents
                            Contents
Other Books by N.D Lewis
Preface
How to get the most from this book
Introduction
Deep Neural Networks
Elman Neural Networks
Jordan Neural Networks
The Secret to the Autoencoder
The Stacked Autoencoder in a Nutshell
Restricted Boltzmann Machines
Deep Belief Networks
A Special Message for YOU
Index
Enjoy these books
                        
Document Text Contents
Page 1

DEEP LEARNING
MADE EASY

WITH R
A Gentle Introduction for Data Science.

Dr. N.D. Lewis

Page 2

Copyright © 2016 by N.D. Lewis

All rights reserved. No part of this publication may be reproduced, dis-
tributed, or transmitted in any form or by any means, including photo-
copying, recording, or other electronic or mechanical methods, without
the prior written permission of the author, except in the case of brief quo-
tations embodied in critical reviews and certain other noncommercial uses
permitted by copyright law. For permission requests, contact the author
at: www.AusCov.com.

Disclaimer: Although the author and publisher have made every effort to
ensure that the information in this book was correct at press time, the
author and publisher do not assume and hereby disclaim any liability to
any party for any loss, damage, or disruption caused by errors or omissions,
whether such errors or omissions result from negligence, accident, or any
other cause.

Ordering Information: Quantity sales. Special discounts are available on
quantity purchases by corporations, associations, and others. For details,
email: [email protected]

Image photography by Deanna Lewis

ISBN: 978-1519514219
ISBN: 1519514212

Page 125

CHAPTER 4. JORDAN NEURAL NETWORKS

How to Classify Protein-Protein interaction
Protein-protein interaction refers to the biological functions
carried out by the proteins within the cell by interacting with
other proteins in other cells as a result of biochemical events
and/or electrostatic forces. Such interactions are believed to
be important in understanding disease pathogenesis and de-
veloping new therapeutic approaches. A number of different
perspectives have been used to study these interactions rang-
ing from biochemistry, quantum chemistry, molecular dynam-
ics, signal transduction, among others78 All this information
enables the creation of large protein interaction databases.

Computer scientists Dilpreet and Singh 79 apply Jordan
neural networks to classify protein-protein interactions. The
sample used in their analysis was derived from three existing
databases80. It contained 753 positive patterns and 656 neg-
ative patterns81. Using amino acid composition of proteins as
input to the Jordan network to classify the percentage of in-
teracting and non-interacting proteins the researchers report a
classification accuracy of 97.25%.

Deep Learning to Woo Spanish Speakers
Neural networks have been successfully applied to the difficult
problem of speech recognition in English82. Accurate classi-
fication of the numerical digits 0 through 9 in Spanish using
Jordan neural networks was investigated by researcher Tellez
Paola83.

In a rather small study, the speech of the ten numerical
digits was recorded with voices from three women and three
men. Each person was requested to repeat each digit four times.
The network was then trained to classify the digits. Using nine
random initializations’ Tellez reports an average classification
accuracy of 96.1%.

109

Page 126

Deep Learning Made Easy with R

Essential Elements for Effective Jor-
dan Models in R

I grew up in the heart of England where the weather is alwayson the move. If you want to experience all four seasons inone day, central England is the place to visit! Anyway, in
England, the weather is always a great conversation starter. So
let’s start our exploration of Jordan networks modeling British
weather. To be specific, we will model the temperature of the
city of Nottingham located in Nottinghamshire, England. You
may recall this area was the hunting ground of the people’s
bandit Robin Hood84. Let’s get our hands dirty and build a
Jordan neural network right now! As with Elman networks,
Jordan networks are great for modeling timeseries data.

Which are the Appropriate Packages?
We will use the RSNNS package along with the quantmod pack-
age. The data frame nottem, in the datasets package, con-
tains monthly measurements on the average air temperature
at Nottingham Castle85, a location Robin Hood would have
known well:
> require(RSNNS)
> data("nottem",package="datasets")
> require(quantmod)

Let’s take a quick peek at the data held in nottem:

110

Page 250

Deep Learning Made Easy with R

Write your notes here:

234

Page 251

INDEX

Write your notes here:

235

Similer Documents