AN EMPIRICAL COMPARISON OF TECHNIQUES FOR HANDLING INCOMPLETE DATA USING DECISION TREES

Increasing the awareness of how incomplete data affects learning and classification accuracy has led to increasing numbers of missing data techniques. This article investigates the robustness and accuracy of seven popular techniques for tolerating incomplete training and test data for different patterns of missing data—different proportions and mechanisms of missing data on resulting tree-based models. The seven missing data techniques were compared by artificially simulating different proportions, patterns, and mechanisms of missing data using 21 complete datasets (i.e., with no missing values) obtained from the University of California, Irvine repository of machine-learning databases (Blake and Merz, 1998 Blake , C. L. and C. J. Merz . 1998 . UCI Repository of machine learning databases . University of California, Department of Information and Computer Science , Irvine , CA . (http:/www.ics.uci.edu/mlearn/MLRepository.html) . [Google Scholar]). A four-way repeated measures design was employed to analyze the data. The simulation results suggest important differences. All methods have their strengths and weaknesses. However, listwise deletion is substantially inferior to the other six techniques, while multiple imputation, that utilizes the expectation maximization algorithm, represents a superior approach to handling incomplete data. Decision tree single imputation and surrogate variables splitting are more severely impacted by missing values distributed among all attributes compared to when they are only on a single attribute. Otherwise, the imputation—versus model-based imputation procedures gave—reasonably good results although some discrepancies remained. Different techniques for addressing missing values when using decision trees can give substantially diverse results, and must be carefully considered to protect against biases and spurious findings. Multiple imputation should always be used, especially if the data contain many missing values. If few values are missing, any of the missing data techniques might be considered. The choice of technique should be guided by the proportion, pattern, and mechanisms of missing data, especially the latter two. However, the use of older techniques like listwise deletion and mean or mode single imputation is no longer justifiable given the accessibility and ease of use of more advanced techniques, such as multiple imputation and supervised learning imputation.     

Caracolomobile: affect in computer systems

This essay presents and reflects upon the construction of a few experimental artworks, among them Caracolomobile, that looks for poetic, aesthetic and functional possibilities to bring computer systems to the sensitive universe of human emotions, feelings and expressions. Modern and Contemporary Art have explored such qualities in unfathomable ways and nowadays is turning towards computer systems and their co-related technologies. This universe characterizes and is the focus of these experimental artworks; artworks dealing with entwined subjective and objective qualities, weaving perceptions, sensations and concepts. One of them, Caracolomobile, features an art installation creating a set up for an artificial robot that recognizes humans’ affective states and answers them with movements and sounds. The robot was installed over an artificial mirror lake in an open indigo-blue space surrounded by mirrors. It perceives and discriminates human emotional states and expressions using an interface developed with a non-intrusive neural headset (The neural headset used was developed by Emotiv Systems: http://www.emotiv.com. Accessed 11 August 2011). This artwork raises questions and looks for answers inquiring about the preliminary steps for the creation of artefacts that would conduct one to poetically experiment with affect, emotion, sensations and feelings in computational systems. Other works in progress ask about the poetic possibilities of mixing computational autonomous processes and behavioural robotic procedures (Arkin 1998) to create artificial environments mixed with humans.     

Activity recognition system using inbuilt sensors of smart mobile phone and minimizing feature vectors

Mobile phone is becoming a very popular tool due to having various user friendly applications with all flexible options. It is highly popular for its light weight, wearable and comfortable uses. Many extrinsic habitat of human being can be monitored by the help of inbuilt sensors and its application software. This has appealing use for healthcare applications using exploitation of Ambient Intelligence for daily activity monitoring system. Here, a standard dataset of UCI HAR (University of California, Irvine, Human Activity Recognition, http://archive.ics.uci.edu) is used for analysis purpose. Naive Bayes Classifier is used for recognition of runtime activities minimizing dimension of large feature vectors. Threshold based condition box is designed by us and finally these two results are compared with that of another classifier HF-SVM (Hardware Friendly-Support Vector Machine) of previous related work.     

An interactive satisficing approach for multi-objective optimization with uncertain parameters

Uncertain variables are used to describe the phenomenon where uncertainty appears in a complex system. For modeling the multi-objective decision-making problems with uncertain parameters, a class of uncertain optimization is suggested for the decision systems in Liu and Chen (2013), http://orsc.edu.cn/online/131020 which is called the uncertain multi-objective programming. In order to solve the proposed uncertain multi-objective programming, an interactive uncertain satisficing approach involving the decision-maker’s flexible demands is proposed in this paper. It makes an improvement in contrast to the noninteractive methods. Finally, a numerical example about the capital budget problem is given to illustrate the effectiveness of the proposed model and the relevant solving approach.     

Data quality,sensitive data and joint controllership as examples of grey areas in the existing data protection framework for the Internet of Things

‘The technology used to provide connectivity from anytime, any place and for anyone. Now anything can be added to the list.’¹¹ITU Internet reports 2005: ‘The Internet of Things’ <http://www.itu.int/dms_pub/itus/opb/pol/S-POL-IR.IT-2005-SUM-PDF-E.pdf> 2, accessed 27 April 2015.View all notes In the Internet of Things (‘IoT’), the amount of individuals’ data collected and processed is increasing substantially as data are being collected from various sources. Most communications between smart devices occur automatically, potentially without the user being aware of it. Many questions arise around the vulnerability of the devices in the IoT, often deployed outside a traditional IT structure and lacking sufficient built-in security. The IoT demands consideration and research into how to best balance the opportunities that the IoT affords against legal risks it imposes on data protection. Considerable questions about how our currently existing EU framework for protection of personal data applies in IoT are being raised. The data protection legislation needs to move from theory to practice and in order to achieve this; the legal framework may need additional mechanisms.     

ELECTRONIC BOOKS IN PUBLIC LIBRARIES: A FEASIBILITY STUDY FOR DEVELOPING USAGE MODELS FOR WEB-BASED AND HARDWARE-BASED ELECTRONIC BOOKS

This article is a slightly abridged edited version of a final report detailing the background and implementation of a project that introduced electronic book (e-book) collections to Essex Public Libraries during 2004. The research considered e-book collections available for borrowing on a PDA (HP iPAQ) and collections downloadable on to the borrower's PDA or PC (OverDrive, ebrary). The project, sponsored by The Laser Foundation,1 The Laser (London and South-Eastern Library Region) Foundation was founded in 2002 following the transfer of an operational business to a grant-making body. The Foundation has made a number of grants to public and academic libraries since 2003, and further detail of its activities can be found at: http://www.bl.uk/concord/laser-about.html (accessed 21 February 2005). consisted of a partnership consisting between Loughborough University,2 The Department of Information Science. http://www.lboro.ac.uk/dis/ (accessed 21 February 2005). Essex Public Libraries3 Essex County Libraries operate 73 public libraries and mobile library services in south-eastern England. Their website can be found at: http://194.129.26.30/vip8/ecc/ECCWebsite/display/channels/libraries_channel_134343_Enjoying/index.jsp (accessed 21 February 2005). and Co-East.4 The Co-East Partnership manages acquisition and provision of electronic resources for public libraries in eastern England, including management of the popular ‘Ask a Librarian’ service. Their website can be found at: http://www.co-east.net/ (accessed 22 October 2004). In addition to a discussion of the findings of the research, guidelines are offered to other public library authorities considering the adoption of e-book collections and mobile technology. Two articles based on this research have been published elsewhere considering the evaluation of the iPAQ trials (Dearnley et al., ) and the provision and uptake of OverDrive and ebrary (Dearnley et al., ) collections.     

The XTS-AES Disk Encryption Algorithm and the Security of Ciphertext Stealing

Abstract

This paper describes the importance of the XTS-AES encryption mode of operation and concludes with a new proof for the security of ciphertext stealing as used by XTS-AES. The XTS-AES mode is designed for encrypting data stored on hard disks where there is not additional space for an integrity field. Given this lack of space for an integrity field, XTS-AES builds on the security of AES by protecting the storage device from many dictionary and copy/paste attacks. The operation of the XTS mode of AES is defined in the IEEE 1619-2007 standard [3 IEEE Std 1619–2007 . April 18, 2008 . The XTS-AES Tweakable Block Cipher. Institute of Electrical and Electronics Engineers, Inc.  [Google Scholar]], and has been adopted by the U.S. National Institute of Standards and Technology (NIST) as an approved mode of operation under FIPS 140-2 [2 Dworkin , M. December 2009 . NIST SP 800-38E, “Recommendation for Block Cipher Modes of Operation: The XTS-AES Mode for Confidentiality on Storage Devices”.  [Google Scholar]]. XTS-AES builds on the XEX (Xor-Encrypt-Xor) mode originally proposed by Rogaway [8 Rogaway , P. 2004 . Efficient Instantiations of Tweakable Block ciphers and Refinements to Modes OCB and PMAC. Advances in Cryptology–Asiacrypt 2004, Lecture Notes in Computer Science, vol. 3329, Springer-Verlag, pp. 16–31. Available at http://www.cs.ucdavis.edu/rogaway/papers/offsets.pdf (Accessed 6 January 2012) . [Google Scholar]].     

Q 2 LEARNING AND ITS APPLICATION TO CAR MODELLING

In this paper we describe an application of Q ² learning, a recently developed approach to machine learning in numerical domains (?uc et al., 2003 ?uc , D. , D. Vladu?i? , and I. Bratko . 2003 . Qualitatively faithful quantitative prediction . Proceedings of the eighteenth International Joint Conference on Artificial Intelligence , 1052 – 1057 , Acapulco . Morgan Kaufmann . [Google Scholar] 2004 ?uc , D. , D. Vladu?i? , and I. Bratko . 2004 . Qualitatively faithful quantitative prediction . Artificial Intelligence , 158 : 189 – 214 . [CSA] [CROSSREF]  [Google Scholar]) to the automated modelling of a complex, industrially relevant mechanical system – a four wheel suspension and steering system of a car. In this experiment, first a qualitative model of this dynamic system was induced from data, and then this model was reified into a quantitative model. The induced qualitative models enable explanation of relations among the variables in the system and, when reified into quantitative models, enable accurate numerical prediction. Furthermore, the qualitative guidance of the quantitative modelling process leads to predictions that are significantly more accurate than those obtained by state-of-the-art numerical learning methods.     

Benchmarking SQL on MapReduce systems using large astronomy databases

In the era of bigdata, with a massive set of digital information of unprecedented volumes being collected and/or produced in several application domains, it becomes more and more difficult to manage and query large data repositories. In the framework of the PetaSky project (http://com.isima.fr/Petasky), we focus on the problem of managing scientific data in the field of cosmology. The data we consider are those of the LSST project (http://www.lsst.org/). The overall size of the database that will be produced is expected to exceed 60 PB (Lsst data challenge handbook, 2012). In order to evaluate the performances of existing SQL On MapReduce data management systems, we conducted extensive experiments by using data and queries from the area of cosmology. The goal of this work is to report on the ability of such systems to support large scale declarative queries. We mainly investigated the impact of data partitioning, indexing and compression on query execution performances.     

A Motivationally Oriented Approach to Understanding Game Appropriation

Game appropriation is currently not well conceptualized. What literature does exists (Griffiths & Light, 2008 Griffiths, M. and Light, B. 2008. Social networking and digital gaming media convergence: Classification and its consequences for appropriation. Information Systems Frontiers, 10: 447–459. [Crossref], [Web of Science ®] , [Google Scholar]; Lowood, 2005 Lowood, H. 2005. Real-time performance: Machinima and game studies. The International Media and Art Association Journal, 1(3): 10–16.  [Google Scholar]; Postigo, 2008 Postigo, H. 2008. Video game appropriation through modifications. Attitudes concerning intellectual property among modders and fans. Convergence, 14: 59–74.  [Google Scholar]; Stalker, 2005 Stalker, P. J. (2005). Gaming in art. Unpublished master's thesis, University of the Witwatersrand, Johannesburg, South Africa. http://www.selectparks.net/dl/PippaStalker_GamingInArt.pdf (http://www.selectparks.net/dl/PippaStalker_GamingInArt.pdf)  [Google Scholar]) uses the term primarily to denote gamers' practices beyond the designers' original intentions, for instance, game content modifications. This article frames game appropriation in a different manner; unlike existing appropriation models, game appropriation is conceptualized as a motivational process underpinned by three primary factors: game design characteristics, social interaction, and the psychological characteristics of the gamer. The main contribution of this article is the development of the first model of game appropriation, the game appropriation model (GAM). GAM explains the process of digital games' incorporation into gamers' daily practices as well as the nature of their gameplay. Game appropriation recognizes the online–offline continuity; it contributes to understating gameplay as a long-term, dynamic activity, directly interrelated with a gamers' everyday life rather than a set of defined moments of participation.     

Optimal control of the production–inventory system with deteriorating items and tradable emission permits

Using the well-known Arrow and Karlin (1958) Arrow, K.J., and Karlin, S. (1958), ‘Production over Time with Increasing Marginal Costs’, in Studies in the Mathematical Theory of Inventory and Production, eds. K.J. Arrow and S. Karlin, Stanford: Stanford University Press, pp. 61–69. [Google Scholar] dynamic production–inventory model and the model with tradable emission permits which was presented by Dobos (2005 Dobos, I. (2005), ‘The Effects of Emission Trading on Production and Inventories in the Arrow–Karlin Model’, International Journal of Production Economics, 93–94, 301–308.[Crossref], [Web of Science ®] , [Google Scholar], 2007) Dobos I. (2007), ‘Tradable Emission Permits and Production-inventory Strategies of the Firm’, International Journal of Production Economics, 108, 329–333.[Crossref], [Web of Science ®] , [Google Scholar], we develop a model of the production–inventory system with deteriorating items and tradable emission permits. The objective of this paper is to apply the optimal control theory to solve the production–inventory problem with deteriorating items and tradable emission permits, and derive the optimal inventory level and the optimal production rate that minimise the total cost. The results are discussed with a numerical example and a sensitivity analysis of the optimal solution with respect to the parameters of the production–inventory system is carried out.     

Autonomous trucks: an affront to masculinity?

Autonomous (or self-driving) vehicle technology has become a commercial reality in a number of cars being driven on public roads around the globe. However, this technology is not only confined to cars. It is equally applicable to trucks and there is speculation that it will be autonomous trucks that will hit the roads first because they operate in ‘a less complicated traffic environment’ and are therefore ideal starting points for the coming automated vehicle market.¹1 https://www.theguardian.com/technology/2015/may/06/nevada-self-driving-trucks-public-roads-daimler-inspiration, last accessed 15 December 2016. For example, in May 2015, authorities granted a license to Daimler to test its self-driving trucks on public roads in the US State of Nevada.²2 https://www.theguardian.com/technology/2015/may/06/nevada-self-driving-trucks-public-roads-daimler-inspiration, last accessed 15 December 2016. These use a series of different technologies, including video cameras and radar, which, when applied in an incremental way, allow drivers to take breaks, although there must always be a qualified driver at the wheel.³3 M Taylor and P Maynard, “Self-driving cars” (2015) 21(5), CTLR 133, 134. It is clear to see the benefits of self-driving trucks to trucking companies, who could use them to transport goods across much longer distances and without all of the scheduled breaks than is presently possible with human truck drivers. But what about the impact on the image of the retained drivers themselves? Traditionally, truck drivers have been men and this representation is well-entrenched in a variety of media. Does the introduction of self-driving truck technology represent an attack on the masculinist truck driving culture as portrayed in TV shows such as ‘Ice Road Truckers’ or ‘Outback Truckers’. If so, is such a change to be encouraged or is the historical image (though objectionable to some) worthy of protection? Given that self-driving trucks could be available for purchase from the middle of the coming decade, the socio-legal aspects of self-driving truck technology will be reflected on and the degree to which this issue can be contemplated at law addressed.     

Work organisation constructs and ergonomic outcomes among European forest machine operators

The aims of the study were to: 1) create work organisation constructs for forest machine operators; 2) examine relationships between these work organisation constructs and different ergonomic outcomes. The study was carried out among forest machine operators in six European countries and was based on a questionnaire survey (n = 358). Exploratory factor analysis (EFA) was used in the formation of the construct measures of work organisation. The EFA resulted in a five-factor structure. The factors were named ‘job control’, ‘variation’, ‘job rotation’, ‘breaks’ and ‘rate of work’. Three multiple regression analyses were carried out separately for the dependent variables job satisfaction, musculoskeletal symptoms and headache/sleeping problems. All regression analyses resulted in significant overall models. The highest explained variance among the dependent variables was found for job satisfaction (R = 0.51, i.e. ‘large’ effect size). The unique contribution of the work organisation constructs varied in the models, but ‘job rotation’ contributed significantly in all regression models. The results may be used as a source for ergonomic improvements to work organisation. The present study was performed as part of the EU project ‘ErgoWood’, which resulted in a European handbook for mechanised forest operations (Gellerstedt et al. 2005 Gellerstedt S. Lidén E. Bohlin F Health and performance in mechanised forest operations The Swedish University of Agricultural Sciences, Department of Forest Products and Markets Uppsala 2005  [Google Scholar]). The results delivered part of the evidence base for a toolbox in this handbook regarding changes in work organisation to improve health and performance.     

A Unified User Management System

ABSTRACT

As open source software has gained in popularity throughout the last decades, free operating systems (OSs) such as Linux (Torvalds) and BSD derivatives (i.e., FreeBSD, 2012; NetBSD, 2012 NetBSD Foundation. (2012). The NetBSD project. Available from http://netbsd.org (http://netbsd.org)  [Google Scholar]; OpenBSD, 2012 OpenBSD. (2012). OpenBSD. OpenBSD. Available from http://netbsd.org (http://netbsd.org)  [Google Scholar]) have become more common, not only on datacenters but also on desktop and laptop computers. It is not rare to find computer labs or company offices composed of personal computers that boot more than one operating system. By being able to choose among available OSs, a company's or organization's information technology manager has the freedom to select the right OS for the company's needs, and the decision can be based on technical or financial criteria. This freedom of choice, however, comes with a cost. The administrative complexity of heterogeneous networks is much higher compared to single OS networks, and if the network is large enough so that protocols such as LDAP (Zeilenga, 2006 Zeilenga, K. 2006. Lightweight directory access protocol (LDAP): Technical specification road map et. alTech. rep., RFC 4510, June [Google Scholar]) or Kerberos (Kohl & Neuman, 1993 Kohl, J. and Neuman, C. 1993. The Kerberos network authentication service (v5) Tech. rep., RFC 1510, September[Crossref] , [Google Scholar]) need to be adopted, then the administration burden may become unbearable. Even though some tools exist that make user management of heterogeneous networks more feasible (Tournier, 2006 Tournier, J. (2006). smbldap-tools – summary [Gna!]. In Welcome to Gna! http://gna.org/projects/smbldap-tools (http://gna.org/projects/smbldap-tools)  [Google Scholar]; Chu & Symas Corp., 2005 Chu, H. and Symas Corp. (2005) http://www.openldap.org/devel/cvsweb.cgi/~checkout~/contrib/slapd-modules/smbk5pwd/README (http://www.openldap.org/devel/cvsweb.cgi/~checkout~/contrib/slapd-modules/smbk5pwd/README)  [Google Scholar]), it is not uncommon to use more than one back end for storing user credentials due to OS incompatibilities. In such configurations, the hardest problem to address is credential and account expiration synchronization among the different back ends. This paper demonstrates a platform that tries to mitigate the problem of synchronization by adding an additional, modular, easy to expand layer which is responsible for synchronizing any number of underlying back ends in a secure fashion.     

Aspect identification and ratings inference for hotel reviews

Today, a large volume of hotel reviews is available on many websites, such as TripAdvisor (http://www.tripadvisor.com) and Orbitz (http://www.orbitz.com). A typical review contains an overall rating, several aspect ratings, and review text. The rating is an abstract of review in terms of numerical points. The task of aspect-based opinion summarization is to extract aspect-specific opinions hidden in the reviews which do not have aspect ratings, so that users can quickly digest them without actually reading through them. The task consists of aspect identification and aspect rating inference. Most existing studies cannot utilize aspect ratings which become increasingly abundant on review hosts. In this paper, we propose two topic models which explicitly model aspect ratings as observed variables to improve the performance of aspect rating inference on unrated reviews. The experiment results show that our approaches outperform the existing methods on the data set crawled from TripAdvisor website.     

Dynamically pre-trained deep recurrent neural networks using environmental monitoring data for predicting PM<Subscript>2.5</Subscript>

Fine particulate matter (\(\hbox {PM}_{2.5}\)) has a considerable impact on human health, the environment and climate change. It is estimated that with better predictions, US$9 billion can be saved over a 10-year period in the USA (State of the science fact sheet air quality. http://www.noaa.gov/factsheets/new, 2012). Therefore, it is crucial to keep developing models and systems that can accurately predict the concentration of major air pollutants. In this paper, our target is to predict \(\hbox {PM}_{2.5}\) concentration in Japan using environmental monitoring data obtained from physical sensors with improved accuracy over the currently employed prediction models. To do so, we propose a deep recurrent neural network (DRNN) that is enhanced with a novel pre-training method using auto-encoder especially designed for time series prediction. Additionally, sensors selection is performed within DRNN without harming the accuracy of the predictions by taking advantage of the sparsity found in the network. The numerical experiments show that DRNN with our proposed pre-training method is superior than when using a canonical and a state-of-the-art auto-encoder training method when applied to time series prediction. The experiments confirm that when compared against the \(\hbox {PM}_{2.5}\) prediction system VENUS (National Institute for Environmental Studies. Visual Atmospheric Environment Utility System. http://envgis5.nies.go.jp/osenyosoku/, 2014), our technique improves the accuracy of \(\hbox {PM}_{2.5}\) concentration level predictions that are being reported in Japan.     

Are mobile phone conversations always so annoying? The ‘need-to-listen’ effect re-visited

According to Monk et al. (2004a Monk, A.F., et al., 2004a. Why are mobile phones annoying? Behaviour & Information Technology, 23 (1), 33–42. doi: 10.1080/01449290310001638496[Taylor & Francis Online], [Web of Science ®] , [Google Scholar]. Why are mobile phones annoying? Behaviour and Information Technology, 23 (1), 33–42), mobile phone conversations are annoying to overhear due to an involuntary need-to-listen in order to predict the inaudible half of the conversation. However, previous support for this need-to-listen explanation of annoyance has failed to consider the confound that mobile phone conversations also have less predictable acoustic patterns and has only investigated ‘neutral’ conversations. By staging mobile and face-to-face conversations in public, this study further supports the need-to-listen explanation. By removing the need-to-listen to the content of a mobile conversation through introducing foreign speech, bystanders no longer perceived the conversation as more annoying than a conversation between two co-present individuals, supporting the need-to-listen explanation over unpredictable acoustics. In two further experiments manipulating conversational content (‘neutral’ vs. ‘intriguing’), findings suggest that the need-to-listen to mobile phone conversations is not inherently annoying; it can be annoying or possibly even ‘interesting’ depending on the conversational content.