Models for Efficient Utilization of Resources for Upgrading Android Mobile Technology

Introduction

In today’s rapidly growing industry, it is very essential to build an effective and reliable connectivity to establish an efficient communication among the clients and employees of any business organization. Communication involves sharing of critical information between the user and the organization where security plays a key role in terms of privacy (Auxilia et al., 2020). The customers want to be connected and be able to communicate with any business organization at any time and from anywhere. The demand of being connected 24/7 is only possible due to the evolution of mobile and Internet technology. The availability of newer operating systems in the competing market plays a significant role in improving the mobile technology. Among many other characteristics and applications; one of the distinguishing characteristics offered by mobile operating systems is that the users can get connected to the internet using the wireless service provider of their smart phones which is cost effective as the mobile system is completely wireless leading to benefits like saving of money and space as compared to wired sensor network (Elfouly et al., 2017). Various types of operating systems are available in the market, the most popular being Android. Android has seen various versions starting from version 1.0, then1.5, 1.6, 2.1, 2.2, 2.3, 3.0and so on. Each version has the improvements over the previous one and thus, it is always advisable to go with the latest version. The upgradation of a version to a higher version may be due to a number of reasons, such as identification of some bug in the previous version, change in the customers’ demands, change in technology in the market, etc. We can also see that change is the need of the hour and we as humans must adapt to those changes. However, seeing from the end of developer’s site, we should understand that incorporating a change due to any of the above listed reason resulting in a new version is not at all an easy task. It comes with lots of difficulty and requires huge amount of resources in terms of time, money and manpower. To elaborate on this, let us understand that the development of a software goes through certain stages before it can be deployed. Incorporating a change in any part of the software (due to any of the reasons stated above) may need widespread changes in different parts of the software and thus lots of rework is required (Sharma et al. 2014). Cost and effort of this rework significantly increases with the stages in software development lifecycle (Boehm and Basili 2001).

The authors in this paper aim to work in the direction of reducing the rework and thus, saving of cost and other resources. For this, the three stable versions of Android, viz Android 6.0 (Marshmallow), Android 7.0 (Nougat) and Android 8.0 (Oreo) are analysed. Each version is fairly large in size consisting of a large number of classes. Due to the availability of limited resources, the developers fail to pay equal attention to all the classes, leading to poor quality software. Thus, the main idea revolves around identification of those classes which are more vulnerable to changes in the next software update. The authors have constructed the prediction models which can be used by the developers/designers in the early phases of software development to identify the classes which need focussed attention. Next, we discuss about the correctness or accuracy of the constructed prediction model. Since the developers are relying on these models for identifying the classes, it is very important that the model should be as accurate as possible. The high dimensionality of the dataset is one of the hindrance which may hamper the accuracy of the models. In addition to this, the high index of features (high dimensionality) makes the computation of data an expensive and tedious task (Rattanawadee and Srivihok, 2015). Dimensionality Reduction refers to the process of reducing the number of dimensions of a given data set. This leads to a reduction in the number of variables and utilization of a group of prime variables.