Experience with an extensible workspace for analysis of low‐level usage data

Analysis of low‐level usage data collected in empirical studies of user interaction is well known as a demanding task. Existing techniques for data collection and analysis are either application specific or data‐driven. This paper presents a workspace for data cleaning, transformation and analysis of low‐level usage data that we have developed and reports our experience with it. By its five‐level architecture, the workspace makes a distinction between more general data that typically can be used in initial data analysis and the data answering a specific research question. The workspace was used in four studies and in total 6.5M user actions were collected from 238 participants. The collected data have been proven to be useful for: (i) validating solution times, (ii) validating process conformances, (iii) exploratory studies on program comprehension for understanding use of classes and documents and (iv) testing hypotheses on keystroke latencies. We have found workspace creation to be demanding in time. Particularly demanding were determining the context of actions and dealing with deficiencies. However, once these processes were understood, it was easy to reuse the workspace for different experiments and to extend it to answer new research questions. Based on our experience, we give a set of guidelines that might help in setting up studies, collecting and preparing data. We recommend that designers of data collection instruments add context to each action. Furthermore, we recommend rapid iterations starting early in the process of data preparation and analysis, and covering both general and specific data. Copyright © 2009 John Wiley & Sons, Ltd.     

The use of data-derived label hierarchies in multi-label classification

Instead of traditional (multi-class) learning approaches that assume label independency, multi-label learning approaches must deal with the existing label dependencies and relations. Many approaches try to model these dependencies in the process of learning and integrate them in the final predictive model, without making a clear difference between the learning process and the process of modeling the label dependencies. Also, the label relations incorporated in the learned model are not directly visible and can not be (re)used in conjunction with other learning approaches. In this paper, we investigate the use of label hierarchies in multi-label classification, constructed in a data-driven manner. We first consider flat label sets and construct label hierarchies from the label sets that appear in the annotations of the training data by using a hierarchical clustering approach. The obtained hierarchies are then used in conjunction with hierarchical multi-label classification (HMC) approaches (two local model approaches for HMC, based on SVMs and PCTs, and two global model approaches, based on PCTs for HMC and ensembles thereof). The experimental results reveal that the use of the data-derived label hierarchy can significantly improve the performance of single predictive models in multi-label classification as compared to the use of a flat label set, while this is not preserved for the ensemble models.     

A quantitative analysis of the unit verification perspective on fault distributions in complex software systems: an operational replication

Unit verification, including software inspections and unit tests, is usually the first code verification phase in the software development process. However, principles of unit verification are weakly explored, mostly due to the lack of data, since unit verification data are rarely systematically collected and only a few studies have been published with such data from industry. Therefore, we explore the theory of fault distributions, originating in the quantitative analysis by Fenton and Ohlsson, in the weakly explored context of unit verification in large-scale software development. We conduct a quantitative case study on a sequence of four development projects on consecutive releases of the same complex software product line system for telecommunication exchanges. We replicate the operationalization from earlier studies, analyzed hypotheses related to the Pareto principle of fault distribution, persistence of faults, effects of module size, and quality in terms of fault densities, however, now from the perspective of unit verification. The patterns in unit verification results resemble those of later verification phases, e.g., regarding the Pareto principle, and may thus be used for prediction and planning purposes. Using unit verification results as predictors may improve the quality and efficiency of software verification.     

Semantic integration of enterprise information systems using meta-metadata ontology

This paper proposes a non-domain-specific metadata ontology as a core component in a semantic model-based document management system (DMS), a potential contender towards the enterprise information systems of the next generation. What we developed is the core semantic component of an ontology-driven DMS, providing a robust semantic base for describing documents’ metadata. We also enabled semantic services such as automated semantic translation of metadata from one domain to another. The core semantic base consists of three semantic layers, each one serving a different view of documents’ metadata. The core semantic component’s base layer represents a non-domain-specific metadata ontology founded on ebRIM specification. The main purpose of this ontology is to serve as a meta-metadata ontology for other domain-specific metadata ontologies. The base semantic layer provides a generic metadata view. For the sake of enabling domain-specific views of documents’ metadata, we implemented two domain-specific metadata ontologies, semantically layered on top of ebRIM, serving domain-specific views of the metadata. In order to enable semantic translation of metadata from one domain to another, we established model-to-model mappings between these semantic layers by introducing SWRL rules. Having the semantic translation of metadata automated not only allows for effortless switching between different metadata views, but also opens the door for automating the process of documents long-term archiving. For the case study, we chose judicial domain as a promising ground for improving the efficiency of the judiciary by introducing the semantics in this field.     

The complementary exponential geometric distribution: Model, properties, and a comparison with its counterpart

In this paper, we proposed a new two-parameter lifetime distribution with increasing failure rate, the complementary exponential geometric distribution, which is complementary to the exponential geometric model proposed by Adamidis and Loukas (1998). The new distribution arises on a latent complementary risks scenario, in which the lifetime associated with a particular risk is not observable; rather, we observe only the maximum lifetime value among all risks. The properties of the proposed distribution are discussed, including a formal proof of its probability density function and explicit algebraic formulas for its reliability and failure rate functions, moments, including the mean and variance, variation coefficient, and modal value. The parameter estimation is based on the usual maximum likelihood approach. We report the results of a misspecification simulation study performed in order to assess the extent of misspecification errors when testing the exponential geometric distribution against our complementary one in the presence of different sample size and censoring percentage. The methodology is illustrated on four real datasets; we also make a comparison between both modeling approaches.     

From insect to Internet: Situated control for networked robot teams

Ant-like systems take advantage of agents' situatedness to reduce or eliminate the need for centralized control or global knowledge. This reduces the need for complexity of individuals and leads to robust, scalable systems. Such insect-inspired situated approaches have proven effective both for task performance and task allocation. The desire for general, principled techniques for situated interaction has led us to study the exploitation of abstract situatedness – situatedness in non-physical environments. The port-arbitrated behavior-based control approach provides a well-structured abstract behavior space in which agents can participate in situated interaction. We focus on the problem of role assumption, distributed task allocation in which each agent selects its own task-performing role. This paper details our general, principled Broadcast of Local Eligibility (BLE) technique for role-assumption in such behavior-space-situated systems, and provides experimental results from the CMOMMT target-tracking task. This revised version was published online in August 2006 with corrections to the Cover Date.     

Interactive visual steering--rapid visual prototyping of a common rail injection system

Interactive steering with visualization has been a common goal of the visualization research community for twenty years, but it is rarely ever realized in practice. In this paper we describe a successful realization of a tightly coupled steering loop, integrating new simulation technology and interactive visual analysis in a prototyping environment for automotive industry system design. Due to increasing pressure on car manufacturers to meet new emission regulations, to improve efficiency, and to reduce noise, both simulation and visualization are pushed to their limits. Automotive system components, such as the powertrain system or the injection system have an increasing number of parameters, and new design approaches are required. It is no longer possible to optimize such a system solely based on experience or forward optimization. By coupling interactive visualization with the simulation back-end (computational steering), it is now possible to quickly prototype a new system, starting from a non-optimized initial prototype and the corresponding simulation model. The prototyping continues through the refinement of the simulation model, of the simulation parameters and through trial-and-error attempts to an optimized solution. The ability to early see the first results from a multidimensional simulation space--thousands of simulations are run for a multidimensional variety of input parameters--and to quickly go back into the simulation and request more runs in particular parameter regions of interest significantly improves the prototyping process and provides a deeper understanding of the system behavior. The excellent results which we achieved for the common rail injection system strongly suggest that our approach has a great potential of being generalized to other, similar scenarios.     

Morphometric and some immunohistochemical characteristics of human choroids plexus stroma and psammoma bodies

Psammoma bodies (PBs) are one of many choroids plexus aging changes. The aim of our research was to perform the quantification of PBs' presence in human choroids plexus stroma, as well as to evaluate the characteristics of choroids plexus stroma in cases in which PBs were present. Afterwards, the observations of the histochemical analysis would be confirmed by immunohistochemical analysis. Choroid plexuses of 30 cadavers were used for the histochemical and, choroids plexuses of 15 cadavers in which PBs' presence was confirmed during the histochemical analysis, were used as material for the immunohistochemical analysis. Light microscopy, histochemical, immunohistochemical, and morphometric method were applied during the study. Classification of the cases was performed by cluster analysis. We observed increase of choroids plexus PBs' presence during the aging process. But this increase is not linear. Their presence is the largest in the second cluster that is younger than the third and older than the first. Nuclear morphometric parameters of the stroma in these cases showed that the cellular composition in this cluster is different than in other two and, that contain larger number of lymphoid cells. Immunohistochemical analysis showed PBs' positive reaction on vimentin, CD45R0, and LCA markers, while in their vicinity, as well as inside them, numerous T-cells were observed. So, the presence of CD45R0 and LCA-positive T cells, PBs' positive reaction on the same markers, indirectly connect these cells with PBs' formation process.     

On the Convergence of Temporal-Difference Learning with Linear Function Approximation

The asymptotic properties of temporal-difference learning algorithms with linear function approximation are analyzed in this paper. The analysis is carried out in the context of the approximation of a discounted cost-to-go function associated with an uncontrolled Markov chain with an uncountable finite-dimensional state-space. Under mild conditions, the almost sure convergence of temporal-difference learning algorithms with linear function approximation is established and an upper bound for their asymptotic approximation error is determined. The obtained results are a generalization and extension of the existing results related to the asymptotic behavior of temporal-difference learning. Moreover, they cover cases to which the existing results cannot be applied, while the adopted assumptions seem to be the weakest possible under which the almost sure convergence of temporal-difference learning algorithms is still possible to be demonstrated.