Search

Inference and testing in general point process models such as the Hawkes model is predominantly based on asymptotic approximations for likelihoodbased estimators and tests, as originally developed in Ogata (1978). As an alternative, and to improve finite sample performance, this paper considers bootstrap-based inference for interval estimation and testing. Specifically, for a wide class of point process models we consider a novel bootstrap scheme labeled 'fixed intensity bootstrap' (FIB), where the conditional intensity is kept fixed across bootstrap repetitions. The FIB, which is very simple to implement and fast in practice, naturally extends previous ideas from the bootstrap literature on time series in discrete time, where the so-called 'fixed design' and 'fixed volatility' bootstrap schemes have shown to be particularly useful and effective. We compare the FIB with the classic recursive bootstrap, which is here labeled 'recursive intensity bootstrap' (RIB). In RIB algorithms, the intensity is stochastic in the bootstrap world and implementation of the bootstrap is more involved, due to its sequential structure. For both bootstrap schemes, no asymptotic theory is available; we therefore provide here a new bootstrap (asymptotic) theory, which allows to assess bootstrap validity. We also introduce novel 'nonparametric' FIB and RIB schemes, which are based on resampling time-changed transformations of the original waiting times. We show effectiveness of the different bootstrap schemes in finite samples through a set of detailed Monte Carlo experiments. As far as we are aware, this is the first detailed Monte Carlo study of bootstrap implementations for Hawkes-type processes. Finally, in order to illustrate, we provide applications of the bootstrap to both financial data and social media data.

We introduce the basic concepts of deep learning. We use TensorFlow 2.x, the Google cloud service, and Google Drive interaction to make the concepts come alive with Python coding examples.

We introduce you to TensorFlow input pipelines with the tf.data API, which enables you to build complex input pipelines from simple, reusable pieces. Input pipelines are the lifeblood of any deep learning experiment because learning models expect data in a TensorFlow consumable form. It is very easy to create high-performance pipelines with the tf.data.Dataset abstraction (a component of the tf.data API) because it represents a sequence of elements from a dataset in a simple format.

We introduce TensorFlow datasets (TFDS). We discuss many facets of a TFDS with code examples. We continue with a complete TFDS modeling example.

In the previous chapter, we demonstrated how to work with TFDS objects. In this chapter, we work through two end-to-end deep learning experiments with large and complex TFDS objects. The Fashion-MNIST and beans datasets are small with simple images.

Feedforward neural nets are generally great for classification and regression problems. CNNs are great for complex image classification. But activations for feedforward nets and CNNs flow only in one direction, from the input layers to the output layer. Since signals flow in only one direction, feedforward and convolutional nets are not ideal if patterns in data change over time. So we need a different network architecture to work with data impacted by time.

We introduce advanced transfer learning with code examples based on several transfer learning architectures. The code examples train learning models with these architectures.

We’ve already leveraged RNNs for NLP. In this chapter, we create experiments to forecast with time series data. We use the famous Weather dataset to demonstrate both a univariate and a multivariate example.

We guide you in the creation of augmented data experiments to increase the diversity of a training set by applying random (but realistic) transformations. Data augmentation is very useful for small datasets because deep learning models crave a lot of data to perform well.

In the previous chapter, we showed you how to work with a TFDS. But what if we want to work with another type of dataset? In this chapter, we show you how to work with other types of data with TensorFlow.

GANs are effective at generating crisp synthetic images, but are limited in size to about 64 × 64 pixels. A Progressive Growing GAN is an extension of the GAN that enables training generator models to generate large high-quality images up to about 1024 × 1024 pixels (as of this writing). The approach has proven effective at generating high-quality synthetic faces that are startlingly realistic.

Reinforcement learning (RL) is an area of machine learning that focuses on teaching intelligent agents how to take actions in an environment in order to maximize cumulative reward. Cumulative reward in RL is the sum of all rewards as a function of the number of training steps.

Autoencoders don’t typically work well with images unless they are very small. But convolutional and variational autoencoders work much better than feedforward dense ones with large color images.

With feedforward neural networks, we achieved good training performance with MNIST and Fashion-MNIST datasets. But images in these datasets are simple and centered within the input space that contains them. That is, they are centered within the pixel matrix that holds them. Input space is all the possible inputs to a model.

We work through a complete deep learning example with Python’s TensorFlow 2.x library in the Google Colab cloud service. We also demonstrate how to link your Google Drive with the Colab cloud service.

Transfer learning is the process of creating new learning models by fine-tuning previously trained neural networks. Instead of training a network from scratch, we download a pre-trained open source learning model and fine-tune it for our own purpose. A pre-trained model is one that is created by someone else to solve a similar problem. We can use one of these instead of building our own model. A big advantage is that a pre-trained model has been crafted by experts, so we can be confident that it performs at a high level (in most cases). Another advantage is that we don’t have to have a lot of data to use a pre-trained model.

Generative modeling is an unsupervised learning technique that involves automatically discovering and learning the regularities (or patterns) in input data so that a trained model can generate new examples that plausibly could have been drawn from the original dataset. A popular type of generative model is a generative adversarial network. Generative adversarial networks (GANs) are generative models that create new data instances that resemble the training data.

We introduce you to Tensor Processing Units with code examples. A Tensor Processing Unit (TPU) is an application-specific integrated circuit (ASIC) designed to accelerate ML workloads. The TPUs available in TensorFlow are custom-developed from the ground up by the Google Brain team based on its plethora of experience and leadership in the ML community. Google Brain is a deep learning artificial intelligence (AI) research team at Google who research ways to make machines intelligent for the improvement of people’s lives.

Now that we have practiced tuning low-dimensional (or simple) data, we are ready to experiment tuning high-dimensional (or complex) data sets. Low-dimensional data consists of a limited number of features, whereas high-dimensional data consists of a very high number of features.

We combine the Anaconda distribution with the popular Scikit-Learn library to demonstrate a wide range of supervised machine learning algorithms supplemented with unsupervised learning algorithms where appropriate. With clear examples, all written in Python, we demonstrate how these algorithms work to solve machine learning problems.

Classification from complex data is handled exactly as with simple data. Data is loaded into feature set X and target y. X data is composed of a matrix of vectors where each vector represents a data element and y data is composed of a vector of targets. However, complex data is composed of a high number of features (hundreds to thousands). Such a data set is commonly referred to as one with a high-dimensional feature space. Text data is also complex because each document must be converted into vectors of numerical values suitable for machine learning algorithms.

Working with data details the earliest processes of data science problem solving. The 1st step is to identify the problem, which determines all else that needs to be done. The 2nd step is to gather data. The 3rd step is to wrangle (munge) data, which is critical. Wrangling is getting data into a form that is useful for machine learning and other data science problems. Of course, wrangled data will probably have to be cleaned. The 4th step is to visualize the data. Visualization helps you get to know the data and, hopefully, identify patterns.

Monte Carlo simulation (MCS) applies repeated random sampling (randomness) to obtain numerical results for deterministic problem solving. It is widely used in optimization, numerical integration, and risk-based decision making. Probability and cumulative density functions are statistical measures that apply probability distributions for random variables, and can be used in conjunction with MCS to solve deterministic problem.

Electronic-commerce has become an important channel for conducting business. Researchers as well as market executives are trying to better understand online consumer behavior. One model used by researchers to understand behavior in the information systems field in general is the technology acceptance model (TAM). The TAM variables are perceived usefulness, perceived ease of use, and intentions. In this study, we suggest the extension of the TAM for its application in the E-commerce field. The original TAM will be extended, by adding four predictor variables. The four predictor variables are process satisfaction, outcome satisfaction, expectations, and E-commerce use. In addition, the TAM will be extended by measuring actual behavior as opposed to measuring intentions as a substitute for actual behavior in previous TAM application studies. We suggest measuring actual use variable in terms of four criterion variables, namely, purchase, access number, access total time, and access average time. The extended TAM is expected to better explain actual behavior in E-commerce environments than the original TAM.

Voice recognition technology-enabled devices possess extraordinary growth potential, yet some research indicates that organizations and consumers are resisting their adoption. This study investigates the implementation of a voice recognition device in the United States Navy. Grounded in the social psychology and information systems literature, the researchers adapted instruments and developed a tool to explain technology adoption in this environment. Using factor analysis and structural equation modeling, analysis of data from the 270 participants explained almost 90% of the variance in the model. This research adapts the technology acceptance model by adding elements of the theory of planned behavior, providing researchers and practitioners with a valuable instrument to predict technology adoption.

Workflow research is commonly concerned with optimization, modeling, and dependency. In this research, we however address a more fundamental issue. By modeling humans and machines as agents and making use of a theoretical computer and statecharts, we prove that many workflow problems do not have computer-based solutions. We also demonstrate a sufficient condition under which computers are able to solve these problems. We end by discussing the relationships between our research and Petri Nets, the multi-agent framework in the literature, linear programming and workflow verification.

In this article, we explore the organizational process change dynamic through a theoretical lens of business process reengineering (BPR) models. We review relevant literature related to such models to devise a synthesized model of BPR. The synthesized model facilitates the identification of success factors for BPR. Results from in-depth case-study research add explanatory power to our model. It is hoped that our model and subsequent success factors will offer insights to help organizations effectively manage change and transformation.

This study investigated the fit between contingency variables (e.g. task uncertainty and decentralization) in accounting information systems (AIS); a good fit is assumed to improve user satisfaction. We divided task uncertainty into two dimensions: variability and analyzability. We chose information scope, timeliness, and aggregation as the AIS characteristics to be studied within each dimension. User satisfaction was adopted as a surrogate to measure the performance of AIS. The responses of 126 managers were drawn from a cross-industry sample in Taiwan. Multiple regression models were employed to test the hypotheses. The results of the empirical test led us to two results: (1) broad-scope information (external environment or future-oriented information) promotes user satisfaction in high task variability situations, and (2) in a highly decentralized organization, broad-scope, timely, and aggregated information will also facilitate user satisfaction.

In this paper, we model business workflows using Harel’s statecharts. Mapping to statecharts allows us to systematically identify potential workflow problems. It also allows us to investigate specific properties inherent in actual business workflows. Our research focuses on three desirable properties of active database systems — termination, confluence, and observable determinism. For termination and confluence, we develop algorithms to provide a theoretical lens linking desirable active database system properties to workflow management systems problems. We initially validate our algorithms by mapping business workflows from a case study. Our research thus builds preliminary theory by developing a systematic method for identifying workflow problems.

We embarked on a case study to explore one organization’s experiences with radical change for the purpose of uncovering how they achieved success. The organization we examined was Honeywell Inc. in Phoenix, Arizona, USA. From the interview data, we were able to devise a set of ten lessons to help others transform successfully. Two important lessons stand out above the rest. First, execution of a carefully developed change plan separates the high performers from less successful BPR projects. Second, recognition that dealing with change is difficult and complicated is not enough. Top management should make change management a top priority and communicate the change vision across the organization.

This paper provides a longitudinal view of one organization’s experiences with IT implementation and Business Process Reengineering since 1990. The organization is EuroProducts; a manufacturer of air freshener and related products located in thecountry, in the West of England, EuroProducts has identified data integration and data standardization as critical to leverage increased performance from its materials requirements process flow. As a result, a new MRP system is being introduced to integrate data input from factory floor workers, management, staff, and IS professionals. The goal is to use the new MRP system as a rallying point to facilitate redesign of material requirements work flows. Aspects of innovation and systems theory are introduced to help the authors organize and identify root causes of the problems EuroProducts has had with its IT implementation and reengineering efforts.

It is always a pleasure to read Mayer's writings. His experiments invariably involve some clever aspect, and his interpretations and extrapolations are creative. The paper at hand meets those standards. In reacting, I shall roughly follow his outline-encoding a problem, schema training, strategy training, and component skills-after an obligatory opening section on what a problem really is.

The results of a field study investigating the impact of the extent of use, staff scheduling, and the adoption of quality control mechanisms (QCM) on end-user perceptions of information technology (IT) in a health care environment are presented. A multivariate model was developed and tested using multiple regression analysis on 47 hospital information system (HIS) end-users. Three specific hypotheses are proposed and tested. Data analyses indicate that end user perceptions depend on two factors: the end-user usage of IT, and the adoption of formal quality control mechanisms. Based on the research results, several suggestions are made to improve the likelihood of successful IT implementation in the health care industry.

This tutorial presents the state-of-the-art of adaptation for web interfaces concerning multi-dimensionality and context-awareness. The specific goals include the presentation of: (i) fundamental concepts, as motivations, definitions and relevant context information; (ii) adaptation techniques for web applications, as methods, models, strategies and technologies; (iii) adaptable and adaptive web applications in scientific and commercial aspects.

Executive information systems represent a significant departure from traditional computer-based information systems. Advocates of executive systems claim that their systems offer the leading-edge option for making computers truly effective for supporting key management functions. This paper examines the interaction between key characteristics of executive systems and fundamental features of human thinking drawn from the field of Behavioral Decision Theory. We examine the possibility that the operation of these systems may reinforce or intensify certain biases in human information processing. We present three examples of potential bias-intensification: availability, regression effects, and overconfidence. Next, the paper applies the concept of decisional guidance to develop design implications for executive systems. Lastly, the paper offers avenues for theoretical development and empirical research.

In this paper, we model business workflows using Harel’s statecharts. We demonstrate that mapping to statecharts allows one to systematically identify potential workflow problems. Moreover, it also allows one to investigate specific properties inherent in actual business workflows. Our research focuses on three desirable properties of active database systems --non-termination, non-confluence, and non-observable determinism. As a theoretical lens for non-termination and non-confluence, we develop algorithms linking desirable active database system properties to workflow management systems problems. Preliminary validation of our algorithms is accomplished by mapping business workflows from a case study. Our research thus generates preliminary theory by developing a systematic method for identifying workflow problems. INTRODUCTION Business workflows can be well defined, predictable, and frequently executed. We thereby refer to these as structured business workflows. Such workflows can be automated by machines to reduce clerical tasks and potential human intervention errors. Workflow management systems (WMS) are a tested vehicle to facilitate automation of structured business workflows. WMS, which are new generations of computerized systems, are designed to manage automated parts of business workflows (Brunwin, 1994). By separating workflow definitions from application software, WMS provide process and knowledge independence, much like data independence provided by database management systems. In this research, we use Harel’s statecharts to model structured business workflows (Harel, 1987) for three reasons. First, Harel’s statecharts are used in the Unified Modeling Revisiting Workflow Modeling with Statecharts 235 Copyright © 2004, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. Language (UML) as a means for modeling behavior (Object Management Group, 1999). Since the UML is the standard modeling language of the Object Management Group1, Harel’s statecharts will soon become common. Second, statecharts are easy to understand and they do not have the problem of exponential growth of states that plague ordinary state transition diagrams (Harel, 1988). We shall elaborate on this point in the Related Work section. Third, their semantics are rigorous enough for formal analysis on various aspects of structured business workflows (Harel & Naamad, 1996). Within the framework of statecharts, we will show how to model workflow concepts and present algorithms that determine whether a given business workflow has certain predefined properties. We will then use a case study with Moore BCS (recently recast as Moore Wallace Incorporated) to explore the characteristics of a business workflow. The algorithms we develop in this study will become part of a software design tool that we will develop in the future. RELATED WORK An overview of workflow management using the latest technology can be found in Georgakopoulos, Hornick, and Sheth (1995). Specification and implementation of exceptions in workflow management systems are discussed in Casati, Ceri, Paraboschi, and Pozzi (1999) and workflow evolution in Casati, Ceri, Pernici, and Pozzi (1998). Active database systems have been studied extensively (Paton & Diaz, 1999). Active database systems and workflow management systems are related since both types of systems employ triggers to respond to external and internal events and exceptions. We are interested in three important properties of active database systems in this research, namely termination, confluence, and observable determinism, which are formally defined in Allen, Hellerstein, and Widom (1995). More discussion on active database systems, which includes several research prototypes and commercial products, can be found in Zaniolo (1997). The statemate approach, which uses statecharts in modeling reactive systems, is described in Harel and Politi (1998) and its semantics in Harel and Naamad (1996). By far, the statemate semantics of statecharts is the most rigorous and precise execution semantics defined for statecharts and it has been in use for more than ten years (Harel & Naamad, 1996). Here we point out the most significant aspects of the execution semantics. The reader may consult Harel and Politi (1998) and Harel and Naamad (1996) for details. The behavior of a system described in statemate semantics is a set of possible runs, each representing the responses of the system to a sequence of external stimuli generated by its environment2. A run consists of a series of detailed snapshots of the system’s situation; such a snapshot is called a status. The first in the sequence is the initial status, and each subsequent one is obtained from its predecessor by executing a step (see Figure 1). Figure 1: Status and Step.

The geometry mathematics of the Design Layout Module of the RDS-Professional aircraft design software is described. This includes the modeling mathematics employed for surfaces and local coordinate systems, the use of an expanded version of the RP8A category numbers as a way to couple design and analysis, and the use of this modeling framework to implement automatic geometric redesign from sizing and MDO results.

In this paper we consider the problems of transforming UML models into object-relational databases, which consist of static aspects and dynamic aspects. For the static aspects of the database, we first remove semantically overloaded elements, and then our algorithms generate Nested Normal Form nested tables. For the dynamic aspects of the database, we show how to map statechart diagrams into triggers. We begin our analysis of a statechart that potentially leads to nontermination of execution. We pay attention to the events, conditions, and actions of the transitions of states in the given statechart. Further we investigate the properties of confluence and observable determinism of the generated triggers.

We develop a holistic model of transformation. Model development is grounded in current transformation theory. We embarked on this daunting task because we believe that holism offers an excellent paradigm for managing change that can be of benefit to both academics and practitioners. We initially test our theoretical model by examining how well it maps onto an existing organization. The organization in question is heavily involved in enterprise transformation. It is a multi-billion dollar organization that is considered to be the top performer in its industry. Our general research question is: how do organizations successfully manage transformation?.

Building on a prior case and field study work, the author developed a process improvement model. The proposed model includes four interdependent components-a systematic methodology to guide quality efforts, an environment conducive to change, empowerment mechanisms for the people who do the work, and information technology (IT) to facilitate automation. The model was used to contrast the process improvement approaches between U.S. and Japanese companies. Results included differences between process improvement approaches, management/worker roles, creativity, change visions, and work environments.

Explores the drivers behind successful business process re‐engineering (BPR). Examines Caterpillar Inc, located in Peoria, Illinois, USA, using data collected over a four‐year period, beginning in 1992. Reveals positive outcomes which can be directly tied to the deployment of the business process simplification and improvement (BSP/I) methodology. BPS/I provides a systematic methodology for analysis, design and implementation of re‐engineering principles. In addition, it provides the environment to help its people generate creative solutions to critical business problems. Examines the link between the steps of BPS/I and the theoretical phases of the problem‐solving process. Also explores the link between BPS/I and Couger’s variant model of creative problem solving. Notes creative opportunities available to Caterpillar as a result of the identification of this link.

This case is a study of the application of tree-based solutions to Idaho National Engineering and Environmental Laboratory (INEEL) challenges in the development of a computerized system to meet complex, yet exacting compliance requirements extended to thousands of employees in a largescale organization. We rehearse the history of the project and include information on the theoretical structure of the tree-based solution used. Using primary research documentation, we use a constructivist approach to the issue of subject matter expert empowerment, a major theme of the case. Of particular interest is how the engineer in question was able to modify his work paradigm to incorporate a new role as digital content designer and overseer of the project. Additionally, the study concentrates on the overall effects of the project on other INEEL systems and working environments at the INEEL. Implications of management- and subject matter expert-directed system design projects using tree-based tools are considered with respect to all aspects of enterprise systems development.

In this article is presented a brief review of the work in machine translation of languages, with the conclusion that much more research needs to be done and that it does not appear that human beings will ever be relieved of the necessity of learning languages and of knowing how to use these languages effectively.