University Applied Data Science and Analytics: Neural Networks

This essay delves into the concept of neural networks, explaining them as algorithms that mimic the human brain to identify underlying data relationships. It defines neural networks as interconnected systems of nodes, highlighting their applications in various fields like finance, business analytics, marketing research, fraud detection, and speech/character recognition. The essay outlines the benefits of neural networks, including their ability to model complex non-linear relationships and generalize from initial inputs. Furthermore, it explores different types of neural networks such as feedforward, radial basis function, multilayer perceptron, convolutional neural networks (CNN), and recurrent neural networks (RNN), providing insights into their functionalities and specific uses in prediction and data analysis. The essay concludes by referencing key publications in the field, offering a comprehensive overview of neural networks and their significance in data science.

Running head: APPLIED DATA SCIENCE AND ANALYTICS
APPLIED DATA SCIENCE AND ANALYTICS
Name of the Student
Name of the University
Author Note

1APPLIED DATA SCIENCE AND ANALYTICS
Discussions
The topic of the essay is neural networks. A neural network is a sequence of
algorithms that endeavours to identify the associations those are underlying in the sequence
of the data via some procedures that imitates a way that the brain of the humans operate
(Klambauer et al 2017). The neural networks are defined as a system of neurons and they can
be either artificial or organic.
A neural network includes various layers of nodes those are interconnected. Every
node is a type of perceptron and is same as more than one regression that is linear. The
perceptron nourishes the signal that is generated by more than one regression that is linear
into a kind of function that is activation and it may not be linear.
Neural network are utilised with the applications for the operations of finances,
planning of enterprise, business analytics and many more. The neural networks have added
acceptance in the applications of business that includes solutions of research of marketing
and forecasting, recognition of fraud and assessment of risk (O’Shea, Corgan and Clancy
2016). The neural networks are also used in the recognition of speech, recognition of
characters, face recognition and there are many such applications where neural networks are
utilised widely.
There are different benefits of neural networks and the benefits of the neural networks are
as follows (Prieto et al 2016)
 They have the capability to learn and then model complicated and non-linear relationships
 The neural networks can generalize after they learn from the inputs those are initial and
their relationships

2APPLIED DATA SCIENCE AND ANALYTICS
 It does not impose any kind of restrictions on the variables those are given as input and
the neural networks can model the data better are highly unstable and a variance that is
non-constant.
There are various types of neural networks that are used in prediction (Hu et al 2016).
The kinds of neural networks are as follows:
 Feed forward neural network: In this, the data bypasses via the different nodes of input till
it reaches the output node.
 Radial basis function Neural Network: it deems the distance of any point that is always
comparative to the centre. These types of neural networks have two different layers.
 More than one layer Perceptron: A more than one layer perceptron has more than three
layers. It is utilised in order to categorize the data that cannot be separated linearly. A
neural network is connected fully.
 Convolution Neural Network: A CNN utilises a deviation of the more than one layer
perceptrons. A CNN includes multiple layers of convolution. These layers can be pooled
or interconnected
 Recurrent Neural Network: It is a type of neural network in which result of specific layer
is saved and then it is given back to input. This assists in order to predict the result of the
layer.

3APPLIED DATA SCIENCE AND ANALYTICS
References
Hu, Z., Ma, X., Liu, Z., Hovy, E. and Xing, E., 2016. Harnessing deep neural networks with
logic rules. arXiv preprint arXiv:1603.06318.
Klambauer, G., Unterthiner, T., Mayr, A. and Hochreiter, S., 2017. Self-normalizing neural
networks. In Advances in neural information processing systems (pp. 971-980).
O’Shea, T.J., Corgan, J. and Clancy, T.C., 2016, September. Convolutional radio modulation
recognition networks. In International conference on engineering applications of neural
networks (pp. 213-226). Springer, Cham.
Prieto, A., Prieto, B., Ortigosa, E.M., Ros, E., Pelayo, F., Ortega, J. and Rojas, I., 2016.
Neural networks: An overview of early research, current frameworks and new
challenges. Neurocomputing, 214, pp.242-268.