Neural Networks in Business

8/8/2019 Neural Networks in Business

1/16

Neural Networks In Business

Presented by:

Namit Jain

Navneet Vashishtha


2/16

Neural networks do not perform

miracles. But if used sensibly

they can produce some amazing

results


3/16

History!!!!!

First attempt was made in 1943 by Mc Culloch & Pitts-

Developed models of neural networks based on their basic

understanding of neurology

Rosenblatt in 1958 developed Perceptron Widrow & hoff in 1960 developed ADALINE

In 1969, Minsky & Papert highlighted the limitation of singlelayer perceptron to multilayered systems


4/16

What Are Neural N/W????

Traditionally, the term neural n/w refers to a n/w or circuit of biological neurons; however the modern usage often refers toartificial neural n/w, composed of artificial neurons or nodes.

Biological neural networks are made up of real biological neurons that are

connected or functionally related in the peripheral nervous system orthe central nervous system. In the field of neuroscience, they are oftenidentified as groups of neurons that perform a specific physiologicalfunction in laboratory analysis.

Artificial neural networks are made up of interconnecting artificial neurons(programming constructs that mimic the properties of biological neurons).Artificial neural networks may either be used to gain an understanding ofbiological neural networks, or for solving artificial intelligence problemswithout necessarily creating a model of a real biological system. The real,biological nervous system is highly complex and includes some featuresthat may seem superfluous based on an understanding of artificialnetworks.


5/16

What it does????

Uses non linear statistical data modeling tools

Current computer given step by step instruction

Human analysis

The computer can learn

Optimal solution provided by computer

Robot issues


6/16

Why use neural n/w????

Adaptive learning

Expert of What if analysis

Self-Organization

Real time operation

Fault tolerance via redundant information coding


7/16

What is neural n/w replacing????

The hype of Artificial intelligence had failed to deliverthe expected computing ability for these reasons:

The business community increased its collection of data,

but the expectations for its usefulness in critical businessassessment & decision making was not realized

Failure to extract meaning from complex & imprecise data

Need to replace the conventional approach to definealgorithm to solve problem and make predictions when

there are no known solution Sequential batch processing of data is inadequate for the

speed at which information is needed


8/16

Applications of neural n/w

Sales forecasting

Industrial process control

Customer research

Data validation

Risk management

Target marketing

Medicine Marketing

Credit evaluation


9/16

Telecommunication

Control of sound & vibration

Stock market prediction

Engine management

Particle accelerator beam control

Credit card fraud detection


10/16

Other uses of Neural Applications

Recognition of speaker in communication

Diagnosis of hepatitis

Recovery of telecommunication from faulty software

Interpretation of multi meaning chinese words

Undersea mine detection

Texture analysis

Three dimensional object recognition Hand written word recognition


11/16

Business uses for neural network

Preventing & Detecting fraud

Neural n/w can identify Patterns of behavior that have led

to fraudulent outcomes in the past

When a similar pattern is repeated, the neural n/wsoftware flags the event for investigation

As more patterns are identified and additional fraud

outcomes are confirmed, the neural n/w software will

begin to learn these patterns and gains human likeintuition


12/16

Establish credit worthiness

Current credit rating use a small number of

predetermined factors to establish an individual credit

worthiness Neural n/w are able to look for hidden pattern of

behavior that may not otherwise be evident to determine

credit worthiness

Again, a neural n/w has the ability to learn & improve itsability to predict over time


13/16

Predicting default or bankruptcy

neural n/w have the ability to sort through a largeamount of data, including customer credit and paymenttransactions

Based on this data, the neural n/w can look for pattern ofbehavior that have led to default or bankruptcy outcomein the past

As the neural n/w software continues to scan customer

behavior patterns, the neural n/w can flag a customer a/cwhen a concerning pattern of credit or payment behavioremerges, allowing the company to intervene before thecustomer defaults or files for bankruptcy


14/16

Marketing Applications

Neural n/w can look at customer purchase behavior over

time to identify purchase patterns Tracking customer purchases, the neural n/w software

can predict what the customer is likely to purchase next

This will allow marketers using neural n/w to target

customers who are likely to purchase their products next,

giving them a competitive advantage


15/16

Vendors

Microsoft (Excel)- Neural Tools

NeuNet Pro

Calsci- Brainmaker

Matlab- Neural network Toolbox


16/16

Neural Networks in Business

Documents