Keith Phalp

An investigation of machine learning based prediction systems

Traditionally, researchers have used either o�f-the-shelf models such as COCOMO, or developed local models using statistical techniques such as stepwise regression, to obtain software eff�ort estimates. More recently, attention has turned to a variety of machine learning methods such as artifcial neural networks (ANNs), case-based reasoning (CBR) and rule induction (RI). This paper outlines some comparative research into the use of these three machine learning methods to build software e�ort prediction systems. We briefly describe each method and then apply the techniques to a dataset of 81 software projects derived from a Canadian software house in the late 1980s. We compare the prediction systems in terms of three factors: accuracy, explanatory value and configurability. We show that ANN methods have superior accuracy and that RI methods are least accurate. However, this view is somewhat counteracted by problems with explanatory value and configurability. For example, we found that considerable eff�ort was required to configure the ANN and that this compared very unfavourably with the other techniques, particularly CBR and least squares regression (LSR). We suggest that further work be carried out, both to further explore interaction between the enduser and the prediction system, and also to facilitate configuration, particularly of ANNs.

B-SCP: A requirements analysis framework for validating strategic alignment of organizational IT based on strategy, context, and process

Abstract Ensuring that organizational IT is in alignment with and provides support for an organizations business strategy is critical to business success. Despite this, business strategy and strategic alignment issues are all but ignored in the requirements engineering research literature. We present B-SCP, a requirements engineering framework for organizational IT that directly addresses an organizations business strategy and the alignment of IT requirements with that strategy. B-SCP integrates the three themes of strategy, context, and process using a requirements engineering notation for each theme. We demonstrate a means of cross-referencing and integrating the notations with each other, enabling explicit traceability between business processes and business strategy. In addition, we show a means of defining requirements problem scope as a Jackson problem diagram by applying a business modeling framework. Our approach is illustrated via application to an exemplar. The case example demonstrates the feasibility of B-SCP, and we present a comparison with other approaches.

The four pillars of crowdsourcing: A reference model

Crowdsourcing is an emerging business model where tasks are accomplished by the general public; the crowd. Crowdsourcing has been used in a variety of disciplines, including information systems development, marketing and operationalization. It has been shown to be a successful model in recommendation systems, multimedia design and evaluation, database design, and search engine evaluation. Despite the increasing academic and industrial interest in crowdsourcing, there is still a high degree of diversity in the interpretation and the application of the concept. This paper analyses the literature and deduces a taxonomy of crowdsourcing. The taxonomy is meant to represent the different configurations of crowdsourcing in its main four pillars: the crowdsourcer, the crowd, the crowdsourced task and the crowdsourcing platform. Our outcome will help researchers and developers as a reference model to concretely and precisely state their particular interpretation and configuration of crowdsourcing.

Quantitative analysis of static models of processes

The upstream activities of software development projects are often viewed as both the most important, the least understood, and hence the most problematic. This is particularly noticeable in terms of satisfying customer requirements. Business process modelling is one solution that is being increasingly used in conjunction with traditional software development, often feeding in to requirements and analysis activities. In addition, research in Systems Engineering for Business Process Change,11Systems Engineering for Business Process Change is a UK research programme funded by the Engineering and Physical Sciences Research Council. highlights the importance of modelling business processes in evolving and maintaining legacy systems that support those processes. However, the major use of business process modelling, is to attempt to restructure the business process, in order to improve some given aspect, e.g., cost or time. This restructuring may be seen either as separate activity or as a pre-cursor to the development of systems to support the new or improved process. The analysis of these business models is, therefore, vital to the improvement of the process and the development of supporting software systems. Supporting this analysis is the focus of this paper. Business processes are typically described with static (diagrammatic) models. This paper proposes a quantitative approach to aid analysis and comparison of these models. This is illustrated using the process-modelling notation, Role Activity Diagrams (RADs). We studied 10 prototyping processes across a number of organisations and found that roles of the same type exhibited similar levels of coupling across processes. Where roles did not adhere to tentative threshold values, further investigation revealed unusual circumstances or hidden behaviour. Notably, analysis of the prototyping roles (which exhibited the greatest variation in coupling), found that coupling was highly correlated with the size of the development team and the number of participants. This suggests that prototyping in large projects had a different process to that for small projects and required more mechanisms for communication. We conclude that counts (measures) may be useful in the analysis of static process models.

The CAP framework for business process modelling

Abstract Business process modelling is an area of work that is increasingly used in conjunction with software development. For example, many development methods note the importance of strategic or business modelling, typically as a prerequisite to analysis. In addition, Systems Engineering for Business Process Change suggests the need to model the business process in maintaining and evolving existing (legacy) systems. In order to model business processes, one needs to consider what notations are most suitable, and what methods to adopt. However, the most appropriate notation typically depends on a number of contextual issues, the purpose of the modelling, the audience for the models and so on. Furthermore, this context changes with the progress of the modelling. Hence, the modeller needs guidance about appropriate approaches at different points in the modelling programme. This paper introduces a framework for business process modelling that provides such guidance without prescribing particular notations. This is achieved by describing business process modelling in terms of three iterative and generic categories or phases: Capture, Analysis and Presentation. The paper shows how different kinds of notational approaches can be used within these categories, discussing the choices available to the modeller. The (CAP) framework is generally applicable, and is illustrated both by a simple theoretical example, and by examples from industrial business process modelling.

Improving the quality of use case descriptions: empirical assessment of writing guidelines

Use cases are the main requirements vehicle of the UML and are used widely to specify systems. Hence, the need to write clear and accurate use case descriptions has a significant impact for many practitioners. However, many have pointed to weaknesses in the support offered to those writing use cases, and a number of authors advocate the use of rules in the composition and structuring of use case descriptions. These rules constrain the user, by only allowing certain grammatical constructions, typically guiding the structure or the style of the description For example, the CREWS research project pioneered Use Case Authoring Guidelines, suggesting that the adoption of such guidelines improved resulting use case descriptions. Replication of CREWS studies appeared to confirm the view that use case descriptions were improved through the application of guideline sets, but also noted that learning such rules presented a significant overhead. Hence, a leaner set of guidelines (the CP rules) was developed.This paper describes empirical work to assess the utility of these two sets of writing guidelines (CREWS and CP). In particular, descriptions are assessed against a set of established criteria—a use case quality description checklist, which the authors described in a previous paper.Our findings suggest that the leaner set of guidelines performs at least as well in terms of their ability to produce clear and accurate (comprehensible) descriptions. Hence, that a tractable set of rules may prove applicable to the industrial context, which could lead to effective validation of use cases.

Combining process modelling methods

Abstract This paper examines two modelling paradigms, namely Hoares Communicating Sequential Processes (CSP) and a subset of Role Activity Diagrams (RADs) and shows how they can be combined to give a new approach to process modelling. We examine the two notations by reference to processes from two different business domains. For each domain, we transform a RAD model (by way of methodical mapping) to arrive at an equivalent formal CSP model. The latter is then explored using a stepper, which allows for process simulation by executing the model. The paper suggests that by providing a mapping between these notations we gain the accessibility of a well understood user-facing modelling paradigm, (RADs), whilst retaining the formality of CSP. This provides us not only with the advantages of understandable user-facing models, for process elicitation and presentation, but also gives us the ability to experiment with (by process simulation) the effects of process change.

Assessing the quality of use case descriptions

Use cases have, for some years, been a popular approach to specification, as part of the Unified Modelling Language (UML). However, a number of authors have pointed to weaknesses with the approach, particularly in terms of the support offered to the writer of the use case description. This paper describes a Use Case Description Quality Checklist that acts as a check on the quality of the written description. The checklist is derived from theories of text comprehension, taken from the Discourse Processing community. The checklist approach has a number of benefits. First, the approach can be used to derive, or examine further, use case guidelines. That is, by considering whether such guidelines are likely to result in desirable qualities within the resulting description, one is able to make an informed judgement about the utility of those guidelines. Second, one can test for the desirable quality features in existing descriptions, thus enabling empirical validation. Third, as a minimum, the quality features can themselves be used as a checklist for the examination, and revision, of use case descriptions. To demonstrate applicability, the paper reports upon the use, and success, of the approach on an industrial case study.

RolEnact - Role Based Enactable Models of Business Processes

This paper describes RolEnact: a process-modelling notation used to provide enactable models of process instances. The paper shows how RolEnact models may be produced which are equivalent to role activity diagrams (RADs). This allows the modeller to describe processes in a notation (RADs); which can be understood both by process consultants and process users, whilst retaining the ability to generate enactable process scenarios.

Crowdsourcing: A Taxonomy and Systematic Mapping Study

Abstract Context: Crowdsourcing, or tapping into the power of the crowd for problem solving, has gained ever-increasing attraction since it was first introduced. Crowdsourcing has been used in different disciplines, and it is becoming well-accepted in the marketplace as a new business model which utilizes Human Intelligence Tasks (HITs). Objective: While both academia and industry have extensively delved into different aspects of crowdsourcing, there seems to be no common understanding of what crowdsourcing really means and what core and optional features it has. Also, we still lack information on the kinds and disciplines of studies conducted on crowdsourcing and how they defined it in the context of their application area. This paper will clarify this ambiguity by analysing the distribution and demographics of research in crowdsourcing and extracting taxonomy of the variability and commonality in the constructs defining the concept in the literature. Method: We conduct a systematic mapping study and analyse 113 papers, selected via a formal process, and report and discuss the results. The study is combined by a content analysis process to extract a taxonomy of features describing crowdsourcing. Results: We extract and describe the taxonomy of features which characterize crowdsourcing in its four constituents; the crowd, the crowdsourcer, the crowdsourced task and the crowdsourcing platform. In addition, we report on different mappings between these features and the characteristics of the studied papers. We also analyse the distribution of the research using multiple criteria and draw conclusions. For example, our results show a constantly increasing interest in the area, especially in North America and a significant interest from industry. Also, we illustrate that although crowdsourcing is shown to be useful in a variety of disciplines, the research in the field of computer science still seems to be dominant in investigating it. Conclusions: This study allows forming a clear picture of the research in crowdsourcing and understanding the different features of crowdsourcing and their popularity, what type of research was conducted, where and how and by whom. The study enables researchers and practitioners to estimate the current status of the research in this new field. Our taxonomy of extracted features provides a reference model which could be used to configure crowdsourcing and also define it precisely and make design decisions on which of its variation to adopt.