Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Data Mining Practical Machine Learning Tools and Techniques

物件導向軟體之架構(Object-Oriented Software Construction)探討

Statistical Methods for Health Care Research

The Object Management Group's (OMG) Model Driven Architecture (MDA) strategy envisages a world where models play a more direct role in software production, being amenable to manipulation and transformation by machine. Model Driven Engineering (MDE) is wider in scope than MDA. MDE combines process and analysis with architecture. This article sets out a framework for model driven engineering, which can be used as a point of reference for activity in this area. It proposes an organisation of the modelling 'space' and how to locate models in that space. It discusses different kinds of mappings between models. It explains why process and architecture are tightly connected. It discusses the importance and nature of tools. It identifies the need for defining families of languages and transformations, and for developing techniques for generating/configuring tools from such definitions. It concludes with a call to align metamodelling with formal language engineering techniques.

Model Driven Engineering

Software systems evolve continuously, requiring continuous maintenance and development. Consequently, their architecture tends to degrade with time as it becomes less relevant to new, emerging requirements. Therefore, stability or resilience is a primary criterion for evaluating an architecture. In this paper, we propose a quantitative approach to study the evolution of the architecture of object oriented systems over time. In particular, we represent an architecture as a set of triplets (S, R, T), where S and T represent two classes and R is a relationship linking them. We use these triplets as basic unit to measure the stability of an architecture. We show the applicability and usefulness of our approach by studying the evolution of three open source systems: JFree Chart and Xerces-J and Rhino.

ADvISE: Architectural Decay in Software Evolution

Search-Based Testing (SBT) uses an evolutionary algorithm to generate test cases. Traditionally, a random selection is used to generate an initial population and also, less often, during the evolution process. Such selection is likely to achieve lower coverage than a guided selection. We define two novel concepts: (1) a constrained population generator (CPG) that generates a diversified initial population that satisfies some test target constraints; and (2) a constrained evolution operator (CEO) that evolves test candidates according to some constraints of the test target. Either the CPG or CEO may substantially increase the chance of reaching adequate coverage with less effort. In this paper, we propose an approach that models a relaxed version of the unit under test as a constraint satisfaction problem. Based on this model and the test target, a CPG generates an initial population. Then, an evolutionary algorithm uses a CEO and this population to generate test input leading to the test target being covered. Our approach combines constraint-based testing (CBT) and SBT and overcomes the limitations associated with each of them. Using eToc, an open-source SBT tool, we implement a prototype of this approach. We present the empirical results of applying both CPG or CEO on three open-source programs and show that CPG or CEO improve SBT performance in terms of branch coverage by 11% while reducing computation time.

/pdf/boosting-search-based-testing-by-using-constraint-based-3t7c2t2p9t.pdf

Boosting search based testing by using constraint based testing

Program dependency artifacts such as call graphs help support a number of software engineering tasks such as software mining, program understanding, debugging, feature location, software maintenance and evolution. Java Enterprise Edition (JEE) applications represent a significant part of the recent legacy applications, and we are interested in modernizing them. This modernization involves, among other things, analyzing dependencies between their various components/tiers. JEE applications tend to be multilanguage, rely on JEE container services, and make extensive use of late binding techniques-all of which makes finding such dependencies difficult. In this paper, we describe some of these difficulties and how we addressed them to build a dependency call graph. We developed our tool called DeJEE (Dependencies in JEE) as an Eclipse plug-in. We applied DeJEE on two open-source JEE applications: Java PetStore and JSP Blog. The results show that DeJEE is able to identify different types of JEE dependencies.

Analyzing program dependencies in Java EE applications

Multi-language systems are common nowadays because most of the systems are developed using components written in different programming languages. These systems could arise from three different reasons: (1) to leverage the strengths and take benefits of each language, (2) to reduce the cost by reusing code written in other languages, (3) to include and accommodate legacy code. However, they also introduce additional challenges, including the increase in the complexity and the need for proper interfaces and interactions between the different languages. To address these challenges, the software-engineering research community, as well as the industry, should describe and provide common guidelines, idioms, and patterns to support the development, maintenance, and evolution of these systems. These patterns are an effective means of improving the quality of multi-language systems. They capture good practices to adopt and bad practices to avoid. In order to help to improve the quality of multi-language systems, we analysed open-source systems, developers' documentation, bug reports, and programming language specifications to extract bad practices of multi-language systems usage. We encoded and cataloged these practices in the form of design anti-patterns. We report here six anti-patterns. These results could help not only researchers but also professional developers considering the use of more than one programming language.

Anti-patterns for multi-language systems

There exists an extensive literature on vision science, on the one hand, and on program comprehension, on the other hand. However, these two domains of research have been so far rather disjoint. Indeed, several cognitive theories have been proposed to explain program comprehension. These theories explain the processes taking place in the software engineers’ minds when they understand programs. They explain how software engineers process available information to perform their tasks but not how software engineers acquire this information. Vision science provides explanations on the processes used by people to acquire visual information from their environment. Joining vision science and program comprehension provides a more comprehensive theoretical framework to explain facts on program comprehension, to predict new facts, and to frame experiments. We join theories in vision science and in program comprehension; the resulting theory is consistent with facts on program comprehension and helps in predicting new facts, in devising experiments, and in putting certain program comprehension concepts in perspective.

Yann-Gaël Guéhéneuc

Papers

ADvISE: Architectural Decay in Software Evolution

Boosting search based testing by using constraint based testing

Analyzing program dependencies in Java EE applications

Anti-patterns for multi-language systems

A Theory of Program Comprehension: Joining Vision Science and Program Comprehension