Brief introduction of neural network including- 1. Fitting Tool 2. Clustering data with a self-organising map 3. Pattern Recognition Tool 4. Time Series Toolbox

1. GROUP N0.-7
Rangpal
Prabhat srivastava
Vineet Kumar
Suyash pandey
Vibhav Yadav
NEURAL NETWORK TOOLBOX
GROUP-7
RANGPAL
VINEET KUMAR
PRABHAT SRIVASTAV
SUYASH PANDEY
VIBHAV YADAV
IET LUCKNOW (AKTU)
U. P. 226021

2.  Artificial neural networks (ANNs) or connectionist
systems are computing systems vaguely inspired by
the biological neural networks that constitute animal
brains.
 Such systems "learn" (i.e. progressively improve
performance on) tasks by considering examples,
generally without task-specific programming. For
example, in image recognition, they might learn to
identify images that contain cats by analyzing example
images that have been manually labeled as "cat" or "no
cat" and using the results to identify cats in other
images. They do this without any a priori knowledge
about cats, e.g., that they have fur, tails, whiskers and
cat-like faces. Instead, they evolve their own set of
relevant characteristics from the learning material that
they process.

3. Neural Network Toolbox provides algorithms,
pretrained models, and apps to create, train,
visualize, and simulate both shallow and deep neural
networks. You can perform classification, regression,
clustering, dimensionality reduction, time-series
forecasting, and dynamic system modeling and
control.
Basically, there are 3 different layers in a neural
network :-
 Input Layer (All the inputs are fed in the model
through this layer)
 Hidden Layers (There can be more than one
hidden layers which are used for processing the
inputs received from the input layers)
 Output Layer (The data after processing is made
available at the output layer)

5.  Collect data
 Create the network — Create Neural Network
Object
 Initialize the weights and biases
 Train the network — Neural Network Training
Concepts
 Validate the network
 Use the network

6.  1-fitting app
 2-Pattern recognition app
 3-clustering app
 4-time series app

7.  AEROSPACE
 AUTOMOTIVE
 BANKING
 DEFENCE
 ELECTRONICS
 ENTERTAINMENT
 FINANCIAL
 INDUSTRIAL
 INSURANCE
 MANUFACTURING
 ROBOTICS
 MEDICAL
 SECURITIES
 TELECOMMUNICATION

8.  FUNCTIONS
 nnstart – neural network getting started GUI
 nftool – neural network fitting tool
 net(view) – view neural network
 net.trainparam.epochs – we can choose how
 many epochs we wants
 net.trainparam.goal – maximum performance
 means we want accuracy
 net.trainparam.lr – learning rate
 net = train(net,x,y) – train network

9.  To define a fitting problem for the toolbox,
arrange a set of Q input vectors as columns in a
matrix. Then, arrange another set of Q target
vectors (the correct output vectors for each of
the input vectors) into a second matrix . For
example, you can define the fitting problem for
a Boolean AND gate with four sets of two-
element input vectors and one-element targets
as follows:
 inputs = [0 1 0 1; 0 0 1 1]; targets = [0 0 0 1];

10.  Training –these are presented to the network
dfuring training,and the network is adjusted
according to its error.
 Validation –these are use to measure network
generalization,and to halt training when
generalization stops improving.
 Testing- these are no effect on training and so
provide and independent measure of network
performance during and after training .

11. cluster is a group of objects that belongs to the same class. In other words, similar
objects are grouped in one cluster and dissimilar objects are grouped in another
cluster.
Clustering is the process of making a group of abstract objects into classes of
similar objects.
The notion of a "cluster" cannot be precisely defined, which is one of the reasons
why there are so many clustering algorithms
The notion of a cluster, as found by different algorithms, varies significantly in its
properties. Understanding these "cluster models" is key to understanding the
differences between the various algorithms. Typical cluster models include:

12. Apps
Neural Net Clustering Cluster data by training a self-organizing
maps network
Functions
nnstart Neural network getting started GUI
view View neural network
selforgmap Self-organizing map
train Train neural network
plotsomhits Plot self-organizing map sample hits
plotsomnc Plot self-organizing map neighbor connections
plotsomnd Plot self-organizing map neighbor distances
plotsomplanes Plot self-organizing map weight plane
plotsompos Plot self-organizing map weight positions
plotsomtop Plot self-organizing map topology
genFunction Generate MATLAB function for simulating neural
network

13. #Connectivity models: for example, hierarchical clustering
#Centroid models: for example, the k-means algorithm
#Density models: for
example, DBSCAN and OPTICS defines clusters as
connected dense regions in the data space.
#Subspace Model
#Group Model
#Graph Based Model
#Neural Model
Clusterings can be roughly distinguished as:
#Hard clustering: each object belongs to a cluster or not
#Soft clustering (also: fuzzy clustering): each object belongs
to each cluster to a certain degree (for example, a likelihood
of belonging to the cluster)

14.  Biology, computational biology and
bioinformatics Ex- Transcryptomics, sequence
analysis, human genetic clustering
 Business and marketing Ex- market research,
Grouping of shopping items
 World wide web Ex- social network analysis,
search result grouping
 Computer science Ex- software evolution,
image segmentation, recommender system

15.  In addition to function fitting, neural networks are also good at
recognizing patterns.
 For example, suppose you want to classify a tumor as benign or
malignant, based on uniformity of cell size, clump thickness,
mitosis, etc. You have 699 example cases for which you have 9
items of data and the correct classification as benign or malignant.
 As with function fitting, there are two ways to solve this problem:
 Use the nprtool GUI, as described in Using the Neural Network
Pattern Recognition Tool.
 Use a command-line solution, as described in Using Command-
Line Functions.
 It is generally best to start with the GUI, and then to use the GUI
to automatically generate command-line scripts. Before using
either method, the first step is to define the problem by selecting a
data set. The next section describes the data format.

16.  To define a pattern recognition problem, arrange a set of Q
input vectors as columns in a matrix. Then arrange another
set of Q target vectors so that they indicate the classes to
which the input vectors are assigned (see"Data
Structures" for a detailed description of data formatting for
static and time-series data).
 When there are only two classes; you set each scalar target
value to either 0 or 1, indicating which class the
corresponding input belongs to. For instance, you can define
the two-class exclusive-or classification problem as follows:
 inputs = [0 1 0 1; 0 0 1 1]; targets = [1 0 0 1; 0 1 1 0];
 When inputs are to be classified into N different classes, the
target vectors have N elements. For each target vector, one
element is 1 and the others are 0. For example, the following
lines show how to define a classification problem that
divides the corners of a 5-by-5-by-5 cube into three classes:

17.  The origin (the first input vector) in one class
 The corner farthest from the origin (the last input vector) in
a second class
 All other points in a third class
 inputs = [0 0 0 0 5 5 5 5; 0 0 5 5 0 0 5 5; 0 5 0 5 0 5 0 5]; targets
= [1 0 0 0 0 0 0 0; 0 1 1 1 1 1 1 0; 0 0 0 0 0 0 0 1];
 Classification problems involving only two classes can be
represented using either format. The targets can consist of
either scalar 1/0 elements or two-element vectors, with one
element being 1 and the other element being 0.
 The next section shows how to train a network to recognize
patterns, using the neural network pattern recognition tool
GUI, nprtool. This example uses the cancer data set
provided with the toolbox. This data set consists of 699 nine-
element input vectors and two-element target vectors. There
are two elements in each target vector, because there are two
categories (benign or malignant) associated with each input
vector.

18. Time series toolbox
• What is time series?
Time series is set of data point listed over
time.
• Why time series toolbox?
It is used for dynamic modelling and
prediction problems.

19. open loop NARX
closed loop NARX
It perform 1 step ahead prediction It
perform multistep ahead prediction
It uses actual value of y for prediction It
uses former value of y for prediction

20. • Time series tool allows to solve 3 kinds of non linear
time series

21. Steps for writing time series code for prediction for NARX output.
(Open loop)
1) Define input and target variables
2) Create a Nonlinear autoregressive network
3) NARXNET function is used to create network
4) NARXNET require (a) InputDelay (b) FeedbackDelay (c)
HiddenlayerSize as input
5) When network got created it require data for training
6) Function PREPARETS prepares time series data and allows to
keep original time series data unchanged
7) PREPARETS uses network ,input series ,target series to preapare
data
8) We know that data prepared is used for three purpose in network
(a) 75% data for training (b)15% data for validation (c)15% data
for testing . So our next step is to divide the data for upcoming
process.
9) Then next step is to train the netwok with the help of TRAIN
function.
10) TRAIN functions uses previously prepared data for training
purpose.

22. Steps for writing time series code for prediction for NAR
output. (closed loop)
1) Closed loop code is not completely different from
open loop
2) Network is always created and trained in open form.
3) So first network is created and trained in open form.
4) Then network is closed using the function
CLOSEDLOOP
5) Further preparets is use to prepare the time series
data
6) Then feedbackdata and input data is updated after
delay
7) Finally performance of network is checked