A methodology for the classification of quality of requirements using machine learning techniques

ContextOne of the most important factors in the development of a software project is the quality of their requirements. Erroneous requirements, if not detected early, may cause many serious problems, such as substantial additional costs, failure to meet the expected objectives and delays in delivery dates. For these reasons, great effort must be devoted in requirements engineering to ensure that the project’s requirements results are of high quality. One of the aims of this discipline is the automatic processing of requirements for assessing their quality; this aim, however, results in a complex task because the quality of requirements depends mostly on the interpretation of experts and the necessities and demands of the project at hand.ObjectiveThe objective of this paper is to assess the quality of requirements automatically, emulating the assessment that a quality expert of a project would assess.MethodThe proposed methodology is based on the idea of learning based on standard metrics that represent the characteristics that an expert takes into consideration when deciding on the good or bad quality of requirements. Using machine learning techniques, a classifier is trained with requirements earlier classified by the expert, which then is used for classifying newly provided requirements.ResultsWe present two approaches to represent the methodology with two situations of the problem in function of the requirement corpus learning balancing, obtaining different results in the accuracy and the efficiency in order to evaluate both representations. The paper demonstrates the reliability of the methodology by presenting a case study with requirements provided by the Requirements Working Group of the INCOSE organization.ConclusionsA methodology that evaluates the quality of requirements written in natural language is presented in order to emulate the quality that the expert would provide for new requirements, with 86.1 of average in the accuracy.     

A methodology for the selection of requirements engineering techniques

The complexity of software projects as well as the multidisciplinary nature of requirements engineering (RE) requires developers to carefully select RE techniques and practices during software development. Nevertheless, the selection of RE techniques is usually based on personal preference or existing company practice rather than on characteristics of the project at hand. Furthermore, there is a lack of guidance on which techniques are suitable for a certain project context. So far, only a limited amount of research has been done regarding the selection of RE techniques based on the attributes of the project under development. The few approaches that currently exist for the selection of RE techniques provide only little guidance for the actual selection process. We believe that the evaluation of RE techniques in the context of an application domain and a specific project is of great importance. This paper describes a Methodology for Requirements Engineering Techniques Selection (MRETS) as an approach that helps requirements engineers select suitable RE techniques for the project at hand. The MRETS has three aspects: Firstly, it aids requirements engineers in establishing a link between the attributes of the project and the attributes of RE techniques. Secondly, based on the evaluation schema proposed in our research, MRETS provides an opportunity to analyze RE techniques in detail using clustering. Thirdly, the objective function used in our approach provides an effective decision support mechanism for the selection of RE techniques. This paper makes contributions to RE techniques analysis, the application of RE techniques in practice, RE research, and software engineering in general. The application of the proposed methodology to an industrial project provides preliminary information on the effectiveness of MRETS for the selection of RE techniques.     

Scenario-based requirements analysis techniques for real-time software systems: a comparative evaluation

One of the most critical phases of software engineering is requirements elicitation and analysis. Success in a software project is influenced by the quality of requirements and their associated analysis since their outputs contribute to higher level design and verification decisions. Real-time software systems are event driven and contain temporal and resource limitation constraints. Natural-language-based specification and analysis of such systems are then limited to identifying functional and non-functional elements only. In order to design an architecture, or to be able to test and verify these systems, a comprehensive understanding of dependencies, concurrency, response times, and resource usage are necessary. Scenario-based analysis techniques provide a way to decompose requirements to understand the said attributes of real-time systems. However they are in themselves inadequate for providing support for all real-time attributes. This paper discusses and evaluates the suitability of certain scenario-based models in a real-time software environment and then proposes an approach, called timed automata, that constructs a formalised view of scenarios that generate timed specifications. This approach represents the operational view of scenarios with the support of a formal representation that is needed for real-time systems. Our results indicate that models with notations and semantic support for representing temporal and resource usage of scenario provide a better analysis domain.H. Saiedian is a member of the Information & Telecommunication Technology Center at the University of Kansas. His research was partially supported by a grant from the National Science Foundation (NSF).     

An operational model for structuring the requirements generation process

Product quality is directly related to how well that product meets the customers needs and intents. Therefore, the ability to capture customer requirements correctly and succinctly is paramount. Unfortunately, within most software development frameworks requirements elicitation, recording and evaluation are some of the more ill-defined and least structured activities. To help address such inadequacies, we propose a requirements generation model (RGM) that (a) decomposes the conventional requirements analysis phase into sub-phases which focus and refine requirement generation activities, (b) bounds and structures those activities to promote a more effective generation process, and (c) implements a monitoring methodology to assist in detecting deviations from well-defined procedures intended to support the generation of requirements that meet the customers intent. The RGM incorporates lessons learned from a preliminary study that concentrated on identifying where and how miscommunication and requirements omission occur. An industry study (also reported in this paper) attests to the effectiveness of the RGM. The results of that study indicate that the RGM helps (a) reduce the late discovery of requirements, (b) reduce the slippage in milestone completion dates, and (c) increase customer and management satisfaction levels.     

Improved practical support for large-scale requirements prioritising

An efficient, accurate and practical process for prioritising requirements is of great importance in commercial software developments. This article improves an existing cost-value approach in which stakeholders compare all unique pairs of candidate requirements according to their value and their cost of implementation. Techniques for reducing the required number of comparisons are suggested, thus making the process more efficient. An initial approach for managing requirements interdependencies is proposed. A support tool for the improved process has been developed to make the process more practical in commercial developments. The improved process and its support tool have been applied and evaluated in an industrial project at Ericsson Radio Systems AB. The results indicate a pressing need for mature processes for prioritising requirements, and the work presented here is an important step in that direction.     

Pair-wise comparisons versus planning game partitioning—experiments on requirements prioritisation techniques

The process of selecting the right set of requirements for a product release is dependent on how well the organisation succeeds in prioritising the requirements candidates. This paper describes two consecutive controlled experiments comparing different requirements prioritisation techniques with the objective of understanding differences in time-consumption, ease of use and accuracy. The first experiment evaluates Pair-wise comparisons and a variation of the Planning game. As the Planning game turned out as superior, the second experiment was designed to compare the Planning game to Tool-supported pair-wise comparisons. The results indicate that the manual pair-wise comparisons is the most time-consuming of the techniques, and also the least easy to use. Tool-supported pair-wise comparisons is the fastest technique and it is as easy to use as the Planning game. The techniques do not differ significantly regarding accuracy. Editor: Daniel Berry     

Formal analysis of the Shlaer-Mellor method: Towards a toolkit of formal and informal requirements specification techniques

In this paper, we define a number of tools that we think belong to the core of any toolkit for requirements engineers. The tools are conceptual and hence, they need precise definitions that lay down as exactly as possible what their meaning and possible use is. We argue that this definition can best be achieved by a formal specification of the tool. This means that for each semi-formal requirements engineering tool we should provide a formal specification that precisely specifies its meaning. We argue that this mutually enhances the formal and semi-formal technique: it makes formal techniques more usable and, as we will argue, at the same time simplifies the diagram-based notations.At the same time, we believe that the tools of the requirements engineer should, where possible, resemble the familiar semi-formal specification techniques used in practice today. In order to achieve this, we should search existing requirements specification techniques to look for a common kernel of familiar semi-formal techniques and try to provide a formalisation for these.In this paper we illustrate this approach by a formal analysis of the Shlaer-Mellor method for object-oriented requirements specification. The formal specification language used in this analysis is LCM, a language based on dynamic logic, but similar results would have been achieved by means of another language. We analyse the techniques used in the information model, state model, process model and communication model of the Shlaer-Mellor method, identify ambiguities and redundancies, indicate how these can be eliminated and propose a formalisation of the result. We conclude with a listing of the tools extracted from the Shlaer-Mellor method that we can add to a toolkit that in addition contains LCM as formal specification technique.     

An evaluation of methods for prioritizing software requirements

This article describes an evaluation of six different methods for prioritizing software requirements. Based on the quality requirements for a telephony system, the authors individually used all six methods on separate occasions to prioritize the requirements. The methods were then characterized according to a number of criteria from a user's perspective. We found the analytic hierarchy process to be the most promising method, although it may be problematic to scale-up. In an industrial follow-up study we used the analytic hierarchy process to further investigate its applicability. We found that the process is demanding but worth the effort because of its ability to provide reliable results, promote knowledge transfer and create consensus among project members.     

Information system for image classification based on frequency curve proximity

With the size digital collections are currently reaching, retrieving the best match of a document from large collections by comparing hundreds of tags is a task that involves considerable algorithm complexity, even more so if the number of tags in the collection is not fixed. For these cases, similarity search appears to be the best retrieval method, but there is a lack of techniques suited for these conditions. This work presents a combination of machine learning algorithms put together to find the most similar object of a given one in a set of pre-processed objects based only on their metadata tags. The algorithm represents objects as character frequency curves and is capable of finding relationships between objects without an apparent association. It can also be parallelized using MapReduce strategies to perform the search. This method can be applied to a wide variety of documents with metadata tags. The case-study used in this work to demonstrate the similarity search technique is that of a collection of image objects in JavaScript Object Notation (JSON) containing metadata tags.     

Language independent semantic kernels for short-text classification

Short-text classification is increasingly used in a wide range of applications. However, it still remains a challenging problem due to the insufficient nature of word occurrences in short-text documents, although some recently developed methods which exploit syntactic or semantic information have enhanced performance in short-text classification. The language-dependency problem, however, caused by the heavy use of grammatical tags and lexical databases, is considered the major drawback of the previous methods when they are applied to applications in diverse languages. In this article, we propose a novel kernel, called language independent semantic (LIS) kernel, which is able to effectively compute the similarity between short-text documents without using grammatical tags and lexical databases. From the experiment results on English and Korean datasets, it is shown that the LIS kernel has better performance than several existing kernels.     

A case study validation of a knowledge-based approach for the selection of requirements engineering techniques

Requirements engineering (RE) is a critical phase in the software engineering process and plays a vital role in ensuring the overall quality of a software product. Recent research has shown that industry increasingly recognizes the importance of good RE practices and the use of appropriate RE techniques. However, due to the large number of RE techniques, requirements engineers find it challenging to select suitable techniques for a particular project. Unfortunately, technique selection based on personal experience has limitations with regards to the scope, effectiveness and suitability of the RE techniques for the project at hand. In this paper, a Knowledge-based Approach for the Selection of Requirements Engineering Techniques (KASRET) is proposed that helps during RE techniques selection. This approach has three major features. First, a library of requirements techniques was developed which includes detailed knowledge about RE techniques. Second, KASRET integrates advantages of different knowledge representation schemata and reasoning mechanisms. Thus, KASRET provides mechanisms for the management of knowledge about requirements techniques and support for RE process development. Third, as a major decision support mechanism, an objective function evaluates the overall ability and cost of RE techniques, which is helpful for the selection of RE techniques. This paper makes not only a contribution to RE but also to research and application of knowledge management and decision support in process development. A case study using an industrial project shows the support of KASRET for RE techniques selection.

基于文本分类技术计算中医方剂相似度

中医方剂是中医药信息的核心内容,借助现代化的工具和手段是处理信息中医药研究的新趋势。在数量逾十万的方剂中,许多方剂之间存在着相似性,一些方剂是在基本方的基础上进行增加或修改形成,研究方剂的相似性有重要意义。定义了相似度的概念,利用自然语言处理领域的知识,基于文本分类算法,计算方剂相似度,为研究中医方剂提供有用信息。     

Advanced product planning: a comprehensive process for systemic definition of new product requirements

This paper reports results of research into the definition of requirements for new consumer products––specifically, electro-mechanical products. The research dealt with the derivation of design requirements that are demonstrably aligned with stakeholder needs. The paper describes a comprehensive process that can enable product development teams to deal with statements of product requirements, as originally collected through market research activities, in a systematic and traceable manner from the early, fuzzy front end, stages of the design process. The process described has been based on principles of systems engineering. A case study from its application and evaluation drawn from the power sector is described in this paper. The case study demonstrates how the process can significantly improve product quality planning practices through revision of captured product requirements, analysis of stakeholder requirements and derivation of design requirements. The paper discusses benefits and issues from the use of the process by product development teams, and identifies areas for further research. Finally, the conclusions drawn from the reported research are presented.

A measurement framework for assessing the maturity of requirements engineering process

Because requirements engineering (RE) problems are widely acknowledged as having a major impact on the effectiveness of the software development process, Sommerville et al. have developed a requirements maturity model. However, research has shown that the measurement process within Sommerville’s model is ambiguous, and implementation of his requirements maturity model leads to confusion. Hence, the objective of our research is to propose a new RE maturity measurement framework (REMMF) based on Sommerville’s model and to provide initial validation of REMMF. The main purpose of proposing REMMF is to allow us to more effectively measure the maturity of the RE processes being used within organisations and to assist practitioners in measuring the maturity of their RE processes. In order to evaluate REMMF, two organisations implemented the measurement framework within their IT divisions, provided us with an assessment of their requirements process and gave feedback on the REMMF measurement process. The results show that our measurement framework is clear, easy to use and provides an entry point through which the practitioners can effectively judge the strengths and weakness of their RE processes. When an organisation knows where it is, it can more effectively plan for improvement.

Comparing the performance of metaheuristics for the analysis of multi-stakeholder tradeoffs in requirements optimisation

Context

In requirements engineering, there will be many different stake holders. Often the requirements engineer has to find a set of requirements that reflect the needs of several different stake holders, while remaining within budget.

Objective

This paper introduces an optimisation-based approach to the automated analysis of requirements assignments when multiple stake holders are to be satisfied by a single choice of requirements.

Method

The paper reports on experiments using two different multi-objective evolutionary optimisation algorithms with real world data sets as well as synthetic data sets. This empirical validation includes a statistical analysis of the performance of the two algorithms.

Results

The results reveal that the Two-Archive algorithm outperformed the others in convergence as the scale of problems increase. The paper also shows how both traditional and animated Kiviat diagrams can be used to visualise the tensions between the stake holders’ competing requirements in the presence of increasing budgetary pressure.

Conclusion

This paper presented the concept of internal tensioning among multi-stakeholder in requirements analysis and optimisation for the first time. This analysis may be useful in internal negotiations over budgetary allowance for the project.     

问答系统中问题模式分类与相似度计算方法

基于FAQ库的限定域自动问答系统由于更具实用性而成为自然语言处理领域的研究热点,而问题之间的相似度计算是其中最关键的技术。现有的问句相似度计算技术在处理带有上下文情景描述的问题时效果较差。针对现有技术存在的问题,提出将用户问题分为简洁模式问题（SMQs）和情景模式问题（CMQs）,并提出了基于规则的问题模式分类算法。在此基础上,进一步提出了综合考察情景相似度和问句相似度的情景模式问题（CMQs）相似度计算方法。实验结果表明,问题模式分类算法取得了90%以上的准确率和召回率,情景模式问题相似度计算方法在时间复杂度较低的情况下也取得了74.3%的正确率。     

Understanding the effects of requirements volatility in software engineering by using analytical modeling and software process simulation

This paper introduces an executable system dynamics simulation model developed to help project managers comprehend the complex impacts related to requirements volatility on a software development project. The simulator extends previous research and adds research results from an empirical survey, including over 50 new parameters derived from the associated survey data, to a base model. The paper discusses detailed results from two cases that show significant cost, schedule, and quality impacts as a result of requirements volatility. The simulator can be used as an effective tool to demonstrate the complex set of factor relationships and effects related to requirements volatility.     

Relationship analysis in requirements engineering

This research addresses a major shortcoming in todays requirements analysis techniques—the lack of a rigorous and comprehensive process to explicitly capture the relationship structure of the problem domain. Whereas other analysis techniques lightly address the relationship discovery process, relationship analysis (RA) is a systematic, domain-independent analysis technique focusing exclusively on a domains relationship structure. This paper describes RAs taxonomy of relationship types and corresponding brainstorming questions for eliciting the relationship structure from a domain expert. A preliminary case study analysis of online bookstores using RA as well as a formal experiment have both confirmed RAs effectiveness in helping the analyst produce significantly higher quality requirements. RA should become an invaluable tool for analysts, irrespective of the software engineering approach taken during systems analysis.     

Addressing privacy requirements in system design: the PriS method

A major challenge in the field of software engineering is to make users trust the software that they use in their every day activities for professional or recreational reasons. Trusting software depends on various elements, one of which is the protection of user privacy. Protecting privacy is about complying with user’s desires when it comes to handling personal information. Users’ privacy can also be defined as the right to determine when, how and to what extend information about them is communicated to others. Current research stresses the need for addressing privacy issues during the system design rather than during the system implementation phase. To this end, this paper describes PriS, a security requirements engineering method, which incorporates privacy requirements early in the system development process. PriS considers privacy requirements as organisational goals that need to be satisfied and adopts the use of privacy-process patterns as a way to: (1) describe the effect of privacy requirements on business processes; and (2) facilitate the identification of the system architecture that best supports the privacy-related business processes. In this way, PriS provides a holistic approach from ‘high-level’ goals to ‘privacy-compliant’ IT systems. The PriS way-of-working is formally defined thus, enabling the development of automated tools for assisting its application.