Analysis of Artificial Neural Networks: Characteristics and Uses

This report provides an in-depth analysis of Artificial Neural Networks (ANNs). It begins by explaining the fundamental characteristics of ANNs, including their mathematical models, interconnected processing elements, and ability to learn, recall, and generalize data. The report then details how and why ANNs are used, highlighting their application of different layers of mathematical processing and their ability to perform image processing, character recognition, and decision-making. Furthermore, it explores the value proposition of ANNs in solving business problems, such as credit card fraud detection and licensing through Optical Character Recognition. A comparison between ANNs and logistic regression is also provided, discussing their similarities in supervised machine learning and differences in structure. The report concludes by emphasizing the growing importance and acceptance of ANNs, along with the continuous research and development aimed at enhancing their effectiveness in various business processes.

Running head: ANALYSIS OF ARTIFICIAL NEURAL NETWORKS
ANALYSIS OF ARTIFICIAL NEURAL NETWORKS
Name of student
Name of university
Author’s note:

1ANALYSIS OF ARTIFICIAL NEURAL NETWORKS
Table of Contents
Introduction................................................................................................................................2
Explain the characteristics of artificial neural networks............................................................2
How and Why ANN is used.......................................................................................................3
Value proposition associated with neural networks for solving business problems..................4
Comparison in between neural networks and logistic regression..............................................4
Conclusion..................................................................................................................................5
References:.................................................................................................................................6

2ANALYSIS OF ARTIFICIAL NEURAL NETWORKS
Introduction
Artificial Neural Network is basically based on simulation of the working process of human
brain. This technology have been gaining higher acceptance and hence wise it can be stated
that proper research and development have been carried out in this field (Van Gerven &
Bohte, 2017). This report discusses about the characteristics of Artificial Neural Network.
This report also provides a vivid understanding of the application of the neural network and
hence wise solve the business problems that are present in the operating process. Comparison
in between the neural network as well as logical regression is also made in the report.
Explain the characteristics of artificial neural networks
Some of the characteristics of artificial neural network are the followings (Chen et al., 2017):
 It is a mathematical model which is implemented considering principles of neural
system.
 It contains various processing elements which are interconnected and these processing
elements are referred to as neurons which is are responsible for executing intelligent
operations of applications for which it is designed
 Information that is stored in the neurons are nothing but the weighted linkage of
neurons
 The input signals associated with the processing elements arrive through various
connections and connecting weights of the network as well
 It is capable of learning, recalling and generalizing data from provided data set and
for this it assigns and adjust weights

3ANALYSIS OF ARTIFICIAL NEURAL NETWORKS
 The behavior of the neurons as collective provides comprehensive idea regarding its
computational power, and this information is not provided in any single neuron and
therefore, it does not include any specific information
How and Why ANN is used
How: It have been seen that Artificial Neural Network have been using different
layers of mathematical processing for making proper sense of the information. However, it
have been seen that there is a choice of more than one neurons and hence wise this can be
stated that selection is made amongst millions of neurons that are available (Lee et al., 2017).
These neurons are named as units. This ensures the fact that arrangement is made in a manner
that are arranged in a series of layer. The output have been received from the outside world.
These are the data sets that are targeted and are desired to be studied. The entire process is
considered to be important as most of the neural networks have been connected via one layer.
These connections are then weighed and the data is hence wise passed via the connections.
One of the most important aspect that have been considered in the process is that the
cognitive neuroscientist learnt that each part of the brain is responsible for certain networking
process and in case that section of the brain is also well understood. It have been seen that
different information are stored in different part of the brain. Hence with the help of the
ANN, proper understanding of the brain operating process can be well understood.
Step 1 that is considered in the process is that the dendrites receives the external signals. In
step 2, the external signs are processed with the help of neuron cell body. In the step 3,
processed signal is converted in output signal and hence wise the entire process is transmitted
via Axon. In step 4, output signals are received by the dendrites of the next neuron via the
platform of synapse.

4ANALYSIS OF ARTIFICIAL NEURAL NETWORKS
Why: There are certain aspects that have been making Artificial Neutral Network famous.
The main reason of implementing this technology is because of proper performing of image
processing and hence wise character recognition. Performing proper forecasting has been
another major advantage that is considered in the operational process. Decision making is the
sector that is affected the most. Hence it can be stated steps that are taken for the decision
making processing have been increasing the need of implementation of the Artificial Neural
Network.
Value proposition associated with neural networks for solving business
problems
Application of neural network includes the fact that there have been proper
provisioning of credit card fraud detection. This is one of the major aspect that have been
considered in a manner that have been ensuring resolving issues regarding frauds. Licensing
is yet another aspect that is to be considered in the process (Zoph & Le, 2016). This section
ensures that the operating process uses the likes of Optical Character Recognition that is
utilized by the fax software. Usage of Lexicus 2 corporations have been used and hence wise
steps have been taken as per required statement. This platform helps in proper assessment of
the cursive handwriting and hence wise analyzing same have been made in the process.
Comparison in between neural networks and logistic regression
Logistic regression and neural networks are two popular algorithm for supervised
machine learning. These algorithms are related to classification problems where number of
possibilities is discrete, not continuous value. Therefore, these two algorithms have various
similarities in these aspects as well (Ruder, 2017). Logistic regression might include single
variable classification where sigmoid function and softmax function are preferred for
multiclass-classification problems. However, neural networks has a specific structure which

5ANALYSIS OF ARTIFICIAL NEURAL NETWORKS
include one input layer, at least one hidden layer and one output layer as well. Therefore, in
terms of structure, these two algorithms follow completely different characteristics.
Therefore, these two algorithms have similarities for example these two algorithms
are considered for providing training to the machines for making these machines capable to
execute requirements of different applications. However, in terms of differences, these two
algorithms follow different structures which makes these algorithms different even though
these algorithms are considered for similar requirements as well.
Conclusion
From the above discussion, it can be stated that the Artificial Neural Network have
been gaining higher importance and acceptance have been in a constant growth. Proper
research and development process have been performed in order to increase the effectiveness
of ANN. The characteristics that are stated in the report have been imposing the fact that it
has a direct effect in the business operational process. Proper description of the comparison is
also stated which helps in proper selection of the tool which will be acting beneficial in the
process.

6ANALYSIS OF ARTIFICIAL NEURAL NETWORKS
References:
Chen, M., Challita, U., Saad, W., Yin, C., & Debbah, M. (2017). Machine learning for
wireless networks with artificial intelligence: A tutorial on neural networks. arXiv
preprint arXiv:1710.02913.
Lee, K. Y., Kim, K. H., Kang, J. J., Choi, S. J., Im, Y. S., Lee, Y. D., & Lim, Y. S. (2017).
Comparison and analysis of linear regression and artificial neural
network. International Journal of Applied Engineering Research, 12(20), 9820-9825.
Ruder, S. (2017). An overview of multi-task learning in deep neural networks. arXiv preprint
arXiv:1706.05098.
Van Gerven, M., & Bohte, S. (2017). Artificial neural networks as models of neural
information processing. Frontiers in Computational Neuroscience, 11, 114.
Zoph, B., & Le, Q. V. (2016). Neural architecture search with reinforcement learning. arXiv
preprint arXiv:1611.01578.