DEVELOPMENT OF ANALOGY BASED ESTIMATION METHOD USING USE CASE COMPLEXITY APPROACH FOR SOFTWARE EFFORT AND COST ESTIMATION Case Study On Government Agencies

• Main Author: KURNIAWAN, IMAM
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/22544
• Institution: Institut Teknologi Bandung

IT implementation in government agencies has increasingly become a necessity, in public services, as well as other business services. This emerges many IT projects, particularly in software development. However, software development projects often have high failure rate. The implementations of IT procurement project in Indonesia present various problems. Several cases of corruption in the IT procurement occurred due to unsuitable software prices. One of the important aspects to consider in software development projects is how to estimate cost in software procurement. <br /> <br /> <br /> The accuracy of software cost estimation is a critical task for deciding software investment in private and government organizations. Cost estimation is usually performed by organization based on early requirements specifications. Accurate estimation results are still needed although the estimation is performed at an early stage. For example in government agencies, owner estimated cost (OEC) is the base to establish the highest bid limit and to assess the fairness of the offers, including cost details. OEC values is considered important in procurement, because if the price set too high, then it can potentially be a high cost for the government, whereas if it is lower from the normal price, then there is a potential failure in the procurement processes because no one would be interested. Therefore, it is important to develop an estimation method that is able to represent software complexity at the early stage of procurement. The approach that can be considered at the early stages is the Use Case approach. Use Case Point (UCP) is a method that has been widely used to estimate the size and effort of software projects based on early specification. However, UCP has some limitations on estimating effort that might affect the accuracy of the estimation results. In the other hand, some previous research resulting that the analogy method had better performance than the algorithmic method. However, the previous studies were using LOC and FP parameters which are less suitable for early estimation. <br /> <br /> <br /> In this study, we propose the development of analogy based estimation methods to improve the estimation accuracy. Development is done by building a new dataset using the use case complexity parameter as a feature to get the project similarity. Datasets are collected from 100 historical software projects data that have been built using the UCP method approach. To calculate effort, IDW's interpolation method is proposed by considering the distance from features similarity between projects. Meanwhile, cost components for the project follow the current procurement regulations in Indonesia. <br /> <br /> <br /> Dataset evaluation for effort estimation using the proposed analogy method shows the best results of MMRE of 0.36 and PRED(0.25) of 0.57. MMRE shows the average difference of actual effort and estimated effort, whereas PRED (0.25) shows prediction level with error value smaller or equal to 25%. These results outperform the UCP method algorithmically using the case of the proposed dataset. Furthermore, the cost estimation of the three software projects using the case study conducted on government agencies resulted in an average percentage deviation of 7.37%. According to the result, this approach has succeeded in increasing the accuracy of the estimated effort and can be used to estimate the software development cost in any organization that needs to conduct an early stage, such as government agency. <br /> <br />