A Survey on Deep Learning Techniques Used for Quality Process

A Survey on Deep Learning Techniques Used for Quality Process

Vanyashree Mardi (Alva's Institute of Engineering and Technology, India), Naresh E. (Jain University, India & Ramaiah Institute of Technology, India) and Vijaya Kumar B. P. (Ramaiah Institute of Technology, India)
Copyright: © 2019 |Pages: 22
DOI: 10.4018/978-1-5225-7862-8.ch008

Abstract

In the current era, software development and software quality has become extensively important for implementing the real-world software application, and it will enhance the software functionality. Moreover, early prediction of expected error and fault level in the quality process is critical to the software development process. Deep learning techniques are the most appropriate methods for this problem, and this chapter carries out an extensive systematic survey on a variety of deep learning. These techniques are used in the software quality process along with a hypothesis justification for each of the proposed solutions. The deep learning and machine learning techniques are considered to be the most suitable systems for software quality prediction. Deep learning is a computational model made up of various hidden layers of investigation used to portray of information with the goal that researchers can better understand complex information issues.
Chapter Preview
Top

Introduction

Software quality is the amount of a system or the standard that measures the system performance. There are two types of software quality measurement. First, the internal quality is measured during the process of software development life cycle (SDLC), while the external quality is related to the functionality that could depend on some of the internal quality attributes. The quality models allow for prediction of the external quality attributes as a function of different internal quality attributes. The tasks involved in prediction of quality attributes are:

  • 1.

    Recognition of the internal quality attributes,

  • 2.

    Description of the relationship between the internal and external attributes.

Many research scholars have proposed different models for software quality prediction. The deep learning approach for predicting the software quality is more efficient than any other methodology.

In this survey, we focused on a variety of deep learning methods used for quality prediction process proposed by different investigators. These deep learning can be defined as neural network with a large number of attributes and hidden layers in one of four essential network architectures:

  • 1.

    Convolutional Neural Networks

  • 2.

    Recurrent Neural Networks

  • 3.

    Recursive Neural Networks

  • 4.

    Unsupervised Pre-training Networks

Software quality plays an important role in our day to day life as constantly require better quality product with more functionality. So to detect the product quality we used neural networks which is very efficient method. The neural network will observes the learning features and match into the target output and actual output. Deep learning method solves the problem of quality prediction by providing better quality of software. The Convolutional Neural Network is a kind of Neural Network which is designed from biologically driven model. The research has been found how humans perceive a quality in different layer. It is very prominent for the quality process kind of application. Convolutional Neural Network which learns peer to peer mapping between low-quality and high qualities. This is absolutely achieved via hidden layers. The entire process of software quality is obtained through learning with pre-processing.

Figure 1.

“Convolutional Neural Network (CNN)

978-1-5225-7862-8.ch008.f01

Convolutional Neural Network is designed in such a way that it provides superior quality accuracy. It requires moderate number of layers and filters. Therefore the proposed methods are faster as compared to other methods and also achieve fast rate for online usage. It does not need to solve any optimization problem because it is a feed-forward model. When large amount of dataset are available then improved the quality of network. The Convolutional Neural Network is trained in a supervised manner to generate high quality product using perceptual loss function which is not based on differences between qualities but instead on differences between high level features extracted from pre-trained convolution Neural Network. During training perceptual loss measures the similarity between output product and target quality product. The proposed method is relatively accurate and fast as compared to other method. This method is efficient as it produces high quality software products.

There are three aspects of this survey:

  • 1.

    Convolutional Neural Network be trained with peer to peer mapping between low to high product with modest preprocessing.

  • 2.

    The idea of designing network structure comes from the relation among the deep learning method and traditional method.

  • 3.

    A Deep learning method is useful in computer vision problem of software application and achieves speed and good quality.

Complete Chapter List

Search this Book:
Reset