Effort Estimation

Sakhrelia [3] proposed to estimate effort based on the number of screens of the mobile applications. Anastasiia [8] suggested to use user stories or features –functional pieces that bring value to the user as driver of driver of the mobile application development effort estimation. In order to estimate the cost of the mobile application, the web pages ―Estimate My App‖ maintain by Oozou [9] suggest to ask 10 questions, including: 1. How big is your app? 2. What level of UI would you like? 3. Users & accounts, 4.User generated contents, 5. Dates and locations, 6. Social & engagement, 7. Billing& e-Commerce, 8. Admin, feedback & analytics, 9. External APIs and Integrations, 10. Security. The cost of the application will be derived from the answers to these 10 questions.

2.2.2 Fuzzy Logic in Software Effort Estimation A fuzzy set theoretic model is a modelling construct featuring two main properties [14]: (1) It operates at a level of linguistic terms (fuzzy sets), and (2) it represents and processes uncertainty. Fuzzy logic offers a particularly convenient way to generate a keen mapping between input and output spaces thanks to the natural expression of fuzzy rules. In software development effort estimation, two considerations justify the decision of implementing a fuzzy model: First, it is impossible to develop a precise mathematical model of the domain [14]; second, metrics only produce estimations of the real complexity. Thus, according to the previous assertions, formulating a tiny set of natural rules describing underlying interactions between the software metrics and the effort estimation could effortlessly reveal their intrinsic and wider correlations.

Over the years, researchers have proposed different framework for medical warehouses. Some researchers [17] focussed on data integration framework. They used multi- agent technology to configure data warehouse to integrate and archives information from heterogeneous sources of health care unit. But from health care professionals’ perspective, main focus is always on user interface and visualization of multi-modal images [18]. Hence the system should be designed to find suitable strategies to integrate important characteristics of multiple data sets into one image such that better insights can be provided [19]. Parvin Vajirkar et al. [20] suggested that context aware applications [21] [22] [23] provide better accuracy and precision to the prediction results. They designed a context aware data mining framework for data mining applications for providing accuracy and precision to prediction results based on context aware factors. Similarly, Szirbik et al. [24] used the rational unified process framework in designing the medical warehouse for elderly patient care system. But all these frameworks lack the completeness; they are either domain specific or focussed on one aspect only. Hence these models cannot be used for generic effort estimation process. For an effective effort estimation model, we have proposed a complete medical warehouse architecture that is flexible enough to be used in different scenario. The architecture shown in Figure 1 is a four layer architecture where the bottom most layer take the data from the operational databases.

Software effort estimation assesses the quantity of work required to develop a software project. It is a well known fact that the software industry is unable to give proper an estimate of effort, time and development cost, and this is described in reports in various reports including those from project management consultancy companies through case studies on failed projects, and surveys. Estimation accuracy results reported are sometimes biased towards high inaccuracy, e.g., consultant’s studies providing estimation advice, journalists’ stories on failed projects or software houses which sell estimation tools. It is hard to have a balanced view on this industry’s estimation performance without unbiased information from representative projects and organizations. Scientific surveys in journals and conferences might be sources for unbiased information [1].

There are many equation based effort estimation models like Bailey-Basil Model, Halstead Model, and Walston-Felix Model. Effort and Cost estimation are the major concern of any sort of software industry. They are calculated with the help of Kilo Line Code (KLOC) which are the number of, line of code in software. For example if a software contain 1000 lines then it has 1 kilo line of codes. There are several ways to estimate it with various advantages and disadvantages. We can distinguish them in two parts, one is equation based and another is model based estimation techniques.

Lava Prasad Kafle [1] has conducted a case study of 5 companies and reviewed over 50 papers spanning 30 years’ time period. Based on these studies it was suggested that testing data estimations need to be recorded, revised and historical evidence has to be taken into account for avoiding the underruns or over runs in testing. The study found that practically the companies were using the expert judgement and empirical evidence based models for estimatingthe verification and validation testing cost efforts.The observations which were summarized as main findings of the case study are- (a) The test effort are calculated similarly as Total Project effort estimation using expert judgement method and (b) Estimation error of testing effort seems to correlate closely to the estimation error of Total project.Study concluded saying that the companies can further improve the cost effort estimation errors of the verification and validation testing and projects with detailed analysis of their processes and by taking the help of consultants.

After the identification of all data movements inside each activity diagram, aggregation of measurement results can be performed by counting the summation of total data movements in the overall functional processes inside each class. Functional size of each class represents the summation of all data movements in that class. Effort estimation model can be constructed using cubic regression. Cubic regression can be applied using two variables, COSMIC function point and the actual effort. COSMIC function point represented as CFP and the actual effort defined as the total effort used for developing the Webapp which present in person/hour p/h. According to Mendes stability [9], the estimating effort model should be tested so that the all data in the dataset which have been used must be tested to confirm symptoms of skewness and outliers. The confirmation can be through applying One- sample Kolmogorov–Smirnov (K–S) and Shapiro–Wilk (S- W) statistical tests. If these symptoms exist, this means that variables require to be normalized to look alike a normal distribution. The variables can be normalized through data transformation using the natural log transformation.

This paper present comprehensive analysis of effort estimation mechanism which is required to save from software crisis. It is concluded that without anomaly consideration it is difficult to determine the accurate estimate of cost and effort. To determine actual effort and cost required for software development, anomalies and size both should be the factors. This will increase the time indicated at beginning required from software development but software crisis will be minimal by adopting this model. Implementation of the proposed system is conducted within MATLAB 2017a. The simulation result in the accurate effort estimation facility which is more than the system without considering the environmental variables. Environmental variables considered does not overload or increase the cost of overall project development but only enhances the estimation factor. Result is improved by the margin of 5%. In future, versatility of humans could also be considered for software effort estimation. This can play critical role in minimizing software crisis. Software crisis although can be minimized by considering environmental variables but still versatility factor can be introduced to achieve high degree of accuracy from the considered system. The multi heuristic approach can be used to achieve optimization within the software development process. Multi heuristic approach such as genetic approach along with versatility detection could be considered as future work.

In Figure 7 the comparison of pre and post code effort estimations is given. The model developed has been applied to some student projects and the graph is plotted. In the figure the X axis represents the project number and Y-axis represents the effort. Series1 indicates the pre coding test effort for the proposed model and Series2 represents traditional method pre coding effort estimation, Series3 represents the post coding test effort for the pro- posed model and Series4 represents the traditional me- thod post coding test effort estimation. It is showing a variation of about 8% over large number of projects. Thus it confirms the fact the estimated efforts both in pre and post coding phase have higher accuracy than the conventional models which as shown earlier show large deviation.

In this chapter, we propose a new method to calculate software size from use cases to overcome the limitations of the UCP model [1]. Team productivity was calculated based on five factors illustrated in Section 4.2.3 instead of the eight environmental factors Table (2-6) proposed in the use case point model. This is because Ochodek et al. [2] argued that the number of environmental factors can be reduced without deteriorating the estimation accuracy. Moreover, in this chapter, we introduce project complexity as a factor that affects software effort in Section 4.2.2. Most importantly, in Chapter 3, the requirements stability factor was one of the eight factors that contribute to productivity; however, we found that requirements stability is an essential factor and thus, we introduce it in Section 4.2.4 as one of four factors that affect software effort estimation. Section 4.2.5 shows the effort-size relationship. Sections 4.3 and 4.4 present the non-linear and linear regression models, respectively. Sections 4.5 and 4.6 present a Radial Basis Function Neural Network (RBFNN) and General Regression Neural Network (GRNN) models, respectively. In Section 4.7, we show how software effort can be estimated based on the regression and neural network models. Sections 4.8 and 4.9 present the verification and evaluation of models, respectively. Section 4.10 presents threats to validity and Section 4.11 concludes the chapter.

I declare that this project report entitled “A CHANGE EFFORT ESTIMATION MODEL USING IMPACT ANALYSIS FOR THE SOFTWARE DEVELOPMENT PHASE” is the result of my own research except as cited in the references. The thesis has not been accepted for any degree and is not concurrently submitted in candidature of any other degree.

Abstract. Global Software Development (GSD) is becoming increasingly preva- lent, with software development teams being distributed around the world and work- ing in collaboration with partner companies despite geographic and time differ- ences. The main advantage of GSD which makes it attractive is the greater avail- ability of human resources at lower costs. However, there are several disadvan- tages which are caused by the distance which separates the development teams. Coordination and communication become more difficult when the software devel- opment teams are located in different places, resulting in hidden costs involved in this process. As such, the effort estimation models used for collocated software de- velopment are inadequate for estimation in GSD. Thus, effort estimation in GSD is becoming an important area of research. Many researchers have focused on effort estimation in GSD over the last decade. This paper presents the findings of a sys- tematic review of the related literature by summarizing the hidden costs in GSD, and discussing the open research issues in effort estimation in GSD.

Approaches based on analogy have shown promise in the field of software effort estimation, and its use has increased among researchers in this area [39]. Authors, such as [37], classify analogy-based technique as a ma- chine learning technique. This technique has been ad- vocated as a potential method for efficient effort estima- tion, since it allows modeling the complexity between the effort and the variables included in the context of the software project (e.g. team data, project data), el- ements which have a relationship that is normally not linear. Wen et al.[78] carried out a systematic review of the literature in which they identified eight types of machine learning techniques. The Case-based reason- ing (CBR) and artificial neural networks (ANN) were the most used techniques for estimating effort, repre- senting 37% and 26%, respectively.

MLP: Neural networks (NNs) are a non-linear modeling tech- nique inspired by the functioning of the human brain [4]–[6] and have previously been applied in the context of software effort estimation [7], [8], [9]. We further discuss Multi Layered Perceptrons (MLPs) which are the most commonly used type of NNs that are based upon a network of neurons arranged in an input layer, one or more hidden layers, and an output layer in a strictly feedforward manner. Each neuron processes its inputs and generates one output value via a transfer function, which is transmitted to the neurons in the subsequent layer. The output of hidden neuron i is computed by processing the weighted inputs and its bias term b i as follows:

The objective of this paper is to explore the studies on software estimation accuracy available in the literature and analyze the findings to obtain the answer for the research question. However, over the years, there have been a lot of researches contributed to software effort estimation models. There are many recognized software estimation models or methods. This paper thus proposed to study 3 prominent estimation models: Function Point Analysis, Use Case Point Analysis, and COCOMO Model since they are more well- known and recognized [2], [5], [6]. Because otherwise it may result in there is no report needed available in the literature had the less well-known or recognized models were selected.

Sergio F. Ochoa et. al, 2003a, in their study of web effort estimation, developed expert centric CWADEE model that can predict efforts for Chilean web application within 24-72 hours of their development. Data Web Points(DWP) were used to measure the functional size of web application. Along with five types of DWP, Cost drives from WebMO[3] were used to obtain functional size of web application. The empirical evaluation of model were performed on 22 web projects and the estimated efforts were rated as good, medium or poor. It was reported that 15 applications were good, 5 were medium and 2 poor. Finally this study revealed that better estimates can be achieved when more experience with CWADEE[17].

Software development productivity measurement and cost estimation has been a re- search topic for more than 3 decades [1], [2], [3]. So far, the vast majority of empiri- cal studies have involved data from proprietary software projects [4]. Even though an increasing number of governments, non-governmental organizations and companies seem interested in using, evaluating and contributing to FLOSS, effort estimation models or other measurement-based models are not generally used within FLOSS communities [4]. Indeed, such exploration and quantification of productivity, specif- ically the determination of how a FLOSS community manages and allocates effort around a major release, may help in comparing FLOSS projects both with propri- etary software projects, and also be useful when making comparisons between large FLOSS communities. Furthermore, such productivity modeling can also help to iden- tify a baseline to measure the possible impact of changes in, for example, processes, methods and tools used by FLOSS communities.

Researching through literature creates the impression that ANNs is a viable alternative to other old methods are used in many cases for effort estimation purposes. Along these lines, a survey on the usage of ANNs for effort estimation has been prepared. Wen et al. [10] assessed distinctive machine learning methods based programming effort, estimation models. It was seen by the examination that machine learning-based effort estimation models perform better than non-machine learning models. Idri et al. [11] suffocated henceforth these are not fitting. As a choice, swift programming change surfaced and is seen to be for the most part recognized here we are tackling an enlightening file from IFPUG subjecting into an invigorating effort estimation, when appeared differently in relation to more settled methods like ace supposition, non-algometric model, for web application checks using the story point approach.

offer some significant improvements in accuracy and has the potential to be a valid additional tool for software effort estimation. Genetic Programming is a nonparametric method since it does not make any assumption about the distribution of the data, and derives the equations according only to fitted values. An effort based model is proposed by [4] for estimation of COCOMO model using genetic algorithm. The algorithm considers methodology linearly related to effort. The model estimates the value of parameters of COCOMO model. The performance of developed model is tested on NASA software projects data. The developed model is found effective in accurate effort estimation. A method [1] has been proposed for feature selection and parameters optimization for machine learning regression for software effort estimation. Simulations are carried out using benchmark data sets of software projects, namely, Desharnais [9], NASA [19], COCOMO [8]. The results are compared to those obtained by methods using neural networks, support vector machines, multiple additive regression trees. In all data sets, the simulations have shown that the proposed GA- based method was able to improve the performance of the machine learning methods. The simulations have also demonstrated that the proposed method outperforms some recent methods for software effort estimation.

In this study, an investigation is performed to determine the influence of effective factors on management of effort estimation in software projects. A questionnaire is designed to collect the opinions of developers about the different aspects of software development effort estimation inside the organization. The questionnaire is designed so that it includes the most important parts of prior survey- based studies as well as the latest issues related to effort estimation in software projects. The survey conducted in this research makes a connection between analytical studies and survey-based studies to clarify the current status of organization for management of effort estimation. Respondents comprise of project developers in a large software company. This research tries to suggest a systematic and structured method in which the aspects of effort estimation are cleared for an organization. Finally, it is expected that the results achieved by this survey to be useful for project managers in field of project planning and scheduling.