Technology transfer to developing countries : the case of physical asset management in South Africa

This research investigates technology transfer (TT) to developing countries, with specific reference to South Africa. Particular attention is paid to physical asset management, which includes the maintenance of plant, equipment and facilities. The research is case based, comprising a main case study (the South African electricity utility, Eskom) and four mini-cases. A five level framework adapted from Salami and Reavill (1997) is used as the methodological basis for the formulation of the research questions. This deals with technology selection, and management issues including implementation and maintenance and evaluation and modifications. The findings suggest the Salami and Reavill (1997) framework is a useful guide for TT. The case organisations did not introduce technology for strategic advantage, but to achieve operational efficiencies through cost reduction, higher quality and the ability to meet customer demand. Acquirers favour standardised technologies with which they are familiar. Cost-benefit evaluations have limited use in technology acquisition decisions. Users rely on supplier expertise to compensate for poor education and technical training in South Africa. The impact of political and economic factors is more evident in Eskom than in the mini-cases. Physical asset management follows traditional preventive maintenance practices, with limited use of new maintenance management thinking. Few modifications of the technology or R&D innovations take place. Little use is made of explicit knowledge from computerised maintenance management systems. Low operating and maintenance skills are not conducive to the transfer of high-technology equipment. South African organisations acquire technology as items of plant, equipment and systems, but limited transfer of technology takes place. This suggests that operators and maintainers frequently do not understand the underlying technology, and like workers elsewhere, are not always inclined towards adopting technology in the workplace.

Uncertainty and topographic visualisations

This thesis is a study of low-dimensional visualisation methods for data visualisation under certainty of the input data. It focuses on the two main feed-forward neural network algorithms which are NeuroScale and Generative Topographic Mapping (GTM) by trying to make both algorithms able to accommodate the uncertainty. The two models are shown not to work well under high levels of noise within the data and need to be modified. The modification of both models, NeuroScale and GTM, are verified by using synthetic data to show their ability to accommodate the noise. The thesis is interested in the controversy surrounding the non-uniqueness of predictive gene lists (PGL) of predicting prognosis outcome of breast cancer patients as available in DNA microarray experiments. Many of these studies have ignored the uncertainty issue resulting in random correlations of sparse model selection in high dimensional spaces. The visualisation techniques are used to confirm that the patients involved in such medical studies are intrinsically unclassifiable on the basis of provided PGL evidence. This additional category of ‘unclassifiable’ should be accommodated within medical decision support systems if serious errors and unnecessary adjuvant therapy are to be avoided.

Cues and knowledge structures used by mental-health professionals when making risk assessments

Background: Research into mental-health risks has tended to focus on epidemiological approaches and to consider pieces of evidence in isolation. Less is known about the particular factors and their patterns of occurrence that influence clinicians’ risk judgements in practice. Aims: To identify the cues used by clinicians to make risk judgements and to explore how these combine within clinicians’ psychological representations of suicide, self-harm, self-neglect, and harm to others. Method: Content analysis was applied to semi-structured interviews conducted with 46 practitioners from various mental-health disciplines, using mind maps to represent the hierarchical relationships of data and concepts. Results: Strong consensus between experts meant their knowledge could be integrated into a single hierarchical structure for each risk. This revealed contrasting emphases between data and concepts underpinning risks, including: reflection and forethought for suicide; motivation for self-harm; situation and context for harm to others; and current presentation for self-neglect. Conclusions: Analysis of experts’ risk-assessment knowledge identified influential cues and their relationships to risks. It can inform development of valid risk-screening decision support systems that combine actuarial evidence with clinical expertise.

A method for automatically eliciting node weights in a hierarchical knowledge-based structure for reasoning with uncertainty

Hierarchical knowledge structures are frequently used within clinical decision support systems as part of the model for generating intelligent advice. The nodes in the hierarchy inevitably have varying influence on the decisionmaking processes, which needs to be reflected by parameters. If the model has been elicited from human experts, it is not feasible to ask them to estimate the parameters because there will be so many in even moderately-sized structures. This paper describes how the parameters could be obtained from data instead, using only a small number of cases. The original method [1] is applied to a particular web-based clinical decision support system called GRiST, which uses its hierarchical knowledge to quantify the risks associated with mental-health problems. The knowledge was elicited from multidisciplinary mental-health practitioners but the tree has several thousand nodes, all requiring an estimation of their relative influence on the assessment process. The method described in the paper shows how they can be obtained from about 200 cases instead. It greatly reduces the experts’ elicitation tasks and has the potential for being generalised to similar knowledge-engineering domains where relative weightings of node siblings are part of the parameter space.

Predictive gene lists for breast cancer prognosis: a topographic visualisation study

Background: The controversy surrounding the non-uniqueness of predictive gene lists (PGL) of small selected subsets of genes from very large potential candidates as available in DNA microarray experiments is now widely acknowledged 1. Many of these studies have focused on constructing discriminative semi-parametric models and as such are also subject to the issue of random correlations of sparse model selection in high dimensional spaces. In this work we outline a different approach based around an unsupervised patient-specific nonlinear topographic projection in predictive gene lists. Methods: We construct nonlinear topographic projection maps based on inter-patient gene-list relative dissimilarities. The Neuroscale, the Stochastic Neighbor Embedding(SNE) and the Locally Linear Embedding(LLE) techniques have been used to construct two-dimensional projective visualisation plots of 70 dimensional PGLs per patient, classifiers are also constructed to identify the prognosis indicator of each patient using the resulting projections from those visualisation techniques and investigate whether a-posteriori two prognosis groups are separable on the evidence of the gene lists. A literature-proposed predictive gene list for breast cancer is benchmarked against a separate gene list using the above methods. Generalisation ability is investigated by using the mapping capability of Neuroscale to visualise the follow-up study, but based on the projections derived from the original dataset. Results: The results indicate that small subsets of patient-specific PGLs have insufficient prognostic dissimilarity to permit a distinction between two prognosis patients. Uncertainty and diversity across multiple gene expressions prevents unambiguous or even confident patient grouping. Comparative projections across different PGLs provide similar results. Conclusion: The random correlation effect to an arbitrary outcome induced by small subset selection from very high dimensional interrelated gene expression profiles leads to an outcome with associated uncertainty. This continuum and uncertainty precludes any attempts at constructing discriminative classifiers. However a patient's gene expression profile could possibly be used in treatment planning, based on knowledge of other patients' responses. We conclude that many of the patients involved in such medical studies are intrinsically unclassifiable on the basis of provided PGL evidence. This additional category of 'unclassifiable' should be accommodated within medical decision support systems if serious errors and unnecessary adjuvant therapy are to be avoided.

An investigation into inductive parameter learning in complex hierarchical knowledge structures representing clinical expertise

This dissertation investigates the very important and current problem of modelling human expertise. This is an apparent issue in any computer system emulating human decision making. It is prominent in Clinical Decision Support Systems (CDSS) due to the complexity of the induction process and the vast number of parameters in most cases. Other issues such as human error and missing or incomplete data present further challenges. In this thesis, the Galatean Risk Screening Tool (GRiST) is used as an example of modelling clinical expertise and parameter elicitation. The tool is a mental health clinical record management system with a top layer of decision support capabilities. It is currently being deployed by several NHS mental health trusts across the UK. The aim of the research is to investigate the problem of parameter elicitation by inducing them from real clinical data rather than from the human experts who provided the decision model. The induced parameters provide an insight into both the data relationships and how experts make decisions themselves. The outcomes help further understand human decision making and, in particular, help GRiST provide more accurate emulations of risk judgements. Although the algorithms and methods presented in this dissertation are applied to GRiST, they can be adopted for other human knowledge engineering domains.

Optimal input/output reduction in production processes

While conventional Data Envelopment Analysis (DEA) models set targets for each operational unit, this paper considers the problem of input/output reduction in a centralized decision making environment. The purpose of this paper is to develop an approach to input/output reduction problem that typically occurs in organizations with a centralized decision-making environment. This paper shows that DEA can make an important contribution to this problem and discusses how DEA-based model can be used to determine an optimal input/output reduction plan. An application in banking sector with limitation in IT investment shows the usefulness of the proposed method.

IT as enabler of sustainable farming:an empirical analysis of farmers' adoption decision of precision agriculture technology

Precision agriculture (PA) describes a suite of IT based tools which allow farmers to electronically monitor soil and crop conditions and analyze treatment options. This study tests a model explaining the difficulties of PA technology adoption. The model draws on theories of technology acceptance and diffusion of innovation and is validated using survey data from farms in Canada. Findings highlight the importance of compatibility among PA technology components and the crucial role of farmers' expertise. The model provides the theoretical and empirical basis for developing policies and initiatives to support PA technology adoption.

Handling varying amounts of missing data when classifying mental-health risk levels

One of the main challenges of classifying clinical data is determining how to handle missing features. Most research favours imputing of missing values or neglecting records that include missing data, both of which can degrade accuracy when missing values exceed a certain level. In this research we propose a methodology to handle data sets with a large percentage of missing values and with high variability in which particular data are missing. Feature selection is effected by picking variables sequentially in order of maximum correlation with the dependent variable and minimum correlation with variables already selected. Classification models are generated individually for each test case based on its particular feature set and the matching data values available in the training population. The method was applied to real patients' anonymous mental-health data where the task was to predict the suicide risk judgement clinicians would give for each patient's data, with eleven possible outcome classes: zero to ten, representing no risk to maximum risk. The results compare favourably with alternative methods and have the advantage of ensuring explanations of risk are based only on the data given, not imputed data. This is important for clinical decision support systems using human expertise for modelling and explaining predictions.

Understanding data collection behaviour of mental health practitioners

Failure to detect patients at risk of attempting suicide can result in tragic consequences. Identifying risks earlier and more accurately helps prevent serious incidents occurring and is the objective of the GRiST clinical decision support system (CDSS). One of the problems it faces is high variability in the type and quantity of data submitted for patients, who are assessed in multiple contexts along the care pathway. Although GRiST identifies up to 138 patient cues to collect, only about half of them are relevant for any one patient and their roles may not be for risk evaluation but more for risk management. This paper explores the data collection behaviour of clinicians using GRiST to see whether it can elucidate which variables are important for risk evaluations and when. The GRiST CDSS is based on a cognitive model of human expertise manifested by a sophisticated hierarchical knowledge structure or tree. This structure is used by the GRiST interface to provide top-down controlled access to the patient data. Our research explores relationships between the answers given to these higher-level 'branch' questions to see whether they can help direct assessors to the most important data, depending on the patient profile and assessment context. The outcome is a model for dynamic data collection driven by the knowledge hierarchy. It has potential for improving other clinical decision support systems operating in domains with high dimensional data that are only partially collected and in a variety of combinations.

Primary vendor capabilities in a mediated outsourcing model:can IT service providers leverage crowdsourcing?

Crowdsourcing platforms that attract a large pool of potential workforce allow organizations to reduce permanent staff levels. However managing this "human cloud" requires new management models and skills. Therefore, Information Technology (IT) service providers engaging in crowdsourcing need to develop new capabilities to successfully utilize crowdsourcing in delivering services to their clients. To explore these capabilities we collected qualitative data from focus groups with crowdsourcing leaders at a large multinational technology organization. New capabilities we identified stem from the need of the traditional service provider to assume a "client" role in the crowdsourcing context, while still acting as a "vendor" in providing services to the end-client. This paper expands the research on vendor capabilities and IT outsourcing as well as offers important insights to organizations that are experimenting with, or considering, crowdsourcing. © 2014 Elsevier B.V. All rights reserved.

Generalized Scalarizing Problems GENS and GENSLex of Multicriteria Optimization

* This paper is partially supported by the National Science Fund of Bulgarian Ministry of Education and Science under contract № I–1401\2004 "Interactive Algorithms and Software Systems Supporting Multicriteria Decision Making".

An Interactive Method of Linear Mixed Integer Multicriteria Optimization

The paper describes a learning-oriented interactive method for solving linear mixed integer problems of multicriteria optimization. The method increases the possibilities of the decision maker (DM) to describe his/her local preferences and at the same time it overcomes some computational difficulties, especially in problems of large dimension. The method is realized in an experimental decision support system for finding the solution of linear mixed integer multicriteria optimization problems.

Classification-Based Method of Linear Multicriteria Optimization

The paper describes a classification-based learning-oriented interactive method for solving linear multicriteria optimization problems. The method allows the decision makers describe their preferences with greater flexibility, accuracy and reliability. The method is realized in an experimental software system supporting the solution of multicriteria optimization problems.

Classification of Smoking Cessation Status Using Various Data Mining Methods

AMS Subj. Classification: 62P10, 62H30, 68T01