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

Hawaii International Conference on System Sciences

Исследование влияния пола и психологических характеристик автора на количественные параметры его текста с использованием программы Linguistic Inquiry and Word Count

During the 1980s, U.S. firms that announced stock repurchase programs earned favorable long-run returns. Recently, concerns have been raised regarding the robustness of these findings. This comes at a time of explosive worldwide growth in the adoption of repurchase programs. This study provides out-of-sample evidence for 1,060 Canadian repurchase programs announced between 1989 and 1997. As in the U.S., the Canadian stock market seems to discount the information contained in repurchase announcements. Value stocks announcing repurchase programs have particularly favorable returns. Canadian law requires companies to report how many shares they repurchase on a monthly basis. We find that managers are sensitive to mispricing as completion rates are higher in cases where undervaluation may be a more important factor. Moreover, trades are linked to price movements; managers buy more shares when prices fall and reduce their buying when prices rise.

Stock Repurchases in Canada: Performance and Strategic Trading

Predicting the Big 5 personality traits from digital footprints on social media: A meta-analysis

A variety of approaches have been recently proposed to automatically infer users' personality from their user generated content in social media. Approaches differ in terms of the machine learning algorithms and the feature sets used, type of utilized footprint, and the social media environment used to collect the data. In this paper, we perform a comparative analysis of state-of-the-art computational personality recognition methods on a varied set of social media ground truth data from Facebook, Twitter and YouTube. We answer three questions: (1) Should personality prediction be treated as a multi-label prediction task (i.e., all personality traits of a given user are predicted at once), or should each trait be identified separately? (2) Which predictive features work well across different on-line environments? and (3) What is the decay in accuracy when porting models trained in one social media environment to another?

/pdf/computational-personality-recognition-in-social-media-20odm2xwhw.pdf

Computational personality recognition in social media

Humans are reflective, adaptive, and social. They recall the past, sometimes discuss these recollections with others and become nostalgic; yet, previous research has not examined nostalgia in social media. This paper investigates the expression of nostalgia within Facebook conversations. The specific themes of nostalgic longing and the expression of personal emotions are investigated in two studies. Study 1 examines 375,857 Facebook posts, and Study 2 expands on this by comparing a sample of 10,000 Facebook and 10,000 general (non-nostalgic) posts. Content analyses of these posts reveal significant evidence of nostalgic expressions in Facebook conversations. Cluster analysis reveals newer themes of nostalgic longing related to family, life stories, historical events (presidential elections, man on the moon and Gandhi), spirituality, appreciation of life, romanticism and fun. General posts tend to focus on spontaneous, spur-of-the-moment greetings, emotions, and day of the week. Nostalgic posts, in contrast, are more reflective, more emotional, and frequently include both positive and negative emotions, which is consistent with a deeper, bittersweet character to nostalgia. The concurrent utilization of past- and present-tense words in nostalgic posts suggests that, for some Facebook users, nostalgia helps interpret and navigate present circumstances. The research concludes with theoretical and managerial implications. We identified newer topics of nostalgic longing on Facebook which are more pensive and reflective, indicating deeper levels of cognitive processing than general Facebook posts, which are more "in the moment."Nostalgic reverie on Facebook contains both cognitive reflection and affective reactions.Emotions expressed in Facebook conversations tend toward the positive and often contain mixed emotions, - documenting the bittersweet nature of nostalgia.Nostalgic Facebook posts frequently discuss the past, but also frequently discuss the present. Users frequently compare the past to the present and utilize the past to reflect on the present circumstances of their lives, the nation and society.These findings have managerial implications as well.Companies can segment Facebook users based on their posts and then target these consumers using nostalgia-based behavioral advertising campaigns.

The good old days

This paper describes the submission of the University of Washington's Center for Data Science to the PAN 2014 author profiling task. We examine the predictive quality in terms of age and gender of several sets of features extracted from various genres of online social media. Through comparison, we establish a feature set which maximizes accuracy of gender and age prediction across all genres examined. We report accuracies obtained by two approaches to the multi-label classification problem of predicting both age and gender; a model wherein the multi-label problem is reduced to a single-label problem using powerset transformation, and a chained classifier approach wherein the output of a dedicated classifier for gender is used as input for a classifier for age.

http://ceur-ws.org/Vol-1180/CLEF2014wn-Pan-MarquardtEt2014.pdf

Age and gender identification in social media

Service-oriented computing (SOC) enables the development and design of loosely coupled software components for integration with other software system. Since most legacy system were not designed and developed with services components, current legacy software systems require modernization (reengineered) into a target system made up of a set of loosely coupled services. A methodology for service-oriented software reengineering (SoSR) is proposed for applying SOC to legacy systems. The SoSR methodology, a synthesis of best practices, is architecture-centric, service-oriented, role-specific, and model-driven. It is conceptualized from a three-service-participants model, 4+1 view model, and RACI chart. The SoSR methodology is applied in the modernization of a legacy system, a retail business information systems. The results show that this methodology can help software developers and system integrators in reengineering tightly coupled legacy information systems into service-oriented information systems. By including a business process engine for executing composite services with existing applications and database servers, SOC can affect future information system design, deployment, and integration

/pdf/service-oriented-software-reengineering-sosr-41mrl1zmow.pdf

Service-Oriented Software Reengineering: SoSR

We investigate 239 firms cited in the SEC's Accounting and Auditing Enforcement Releases (AAERs). We document significantly negative abnormal operating performance (measured using both cash-flow-based and earnings-based metrics) in the second and third years following AAERs. We also detect significantly negative abnormal stock returns in up to three years following AAERs. We further find that AAER firms are more likely to fail in the post-AAER period. Taken together, our findings suggest that the negative implications of an AAER citation resulting from egregious financial reporting violations can be long lasting and influence various facets of firm performance and survivability.

Sergio Davalos

Papers

Computational personality recognition in social media

The good old days

Age and gender identification in social media

Service-Oriented Software Reengineering: SoSR

The Long‐Term Performance and Failure Risk of Firms Cited in the US SEC's Accounting and Auditing Enforcement Releases