Abstract:
Technology Intelligence entails an intricate process of gathering and transforming data related to technological advancements into refined actionable intelligence by identifying and analysing emergent characteristics and inter-relational linkages. Specialists in this field process this refined information to develop knowledge critical for directing strategic technology management decisions. As primary sources of technology-related data, technology indicators facilitate an expansive characterisation and evaluation of various technologies throughout their entire lifecycle. Engaging in Future-oriented Technology Analysis necessitates a rigorous examination of in-formation from these indicators, equipping decision-makers with sophisticated insights for Technology Forecasting, a vital tool in anticipating and preparing for future technology trends and developments.
This study posits that one can conceptualise Technology Forecasting as a context-sensitive Data Fusion process using Structural Equation Modelling. To this end, this study undertook inductive reasoning to develop generic frameworks for Structural equation Modelling, context-sensitive Data Fusion and the relational mapping of technology indicators for Technology Forecasting, employing analytic literature assessment, conceptual framework development and grounded theory as supportive research methodologies. These generic frameworks were then integrated through inductive reasoning, incorporating comparative and cross-disciplinary analyses and framework unification to develop the study's proposed framework for Technology Forecasting using Structural Equation Modelling-based, context-sensitive Data Fusion.
The proposed Technology Forecasting framework includes methodologies and processes for integrating data from varied sources while emphasising the importance of complex hierarchal relational interconnections and context-related information to augment technology indicator relevance. Data Fusion methods attuned to context refine the output knowledge produced by accounting for the influence of external, context-related variables. Structural Equation Modelling is a robust statistical methodology that can discern and assess the complex hierarchical relationships between latent and observable variables within a given problem and its context, demonstrating efficacy in executing context-sensitive Data Fusion.
The study developed an autoregression model instantiation of the framework, tailored explicitly for longitudinal forecasting in the National Research and Education Network technology do-main. This model instantiation, formulated through deductive reasoning, incorporates insights from action research within the South African National Research Network. It is supplemented by analysis of secondary data from the Trans-European Research and Education Network As-sociation's Compendiums, which record infrastructure, services, and ecosystem-related trends for National Research and Education Networks in Europe.
This autoregressive model instantiation, although found suboptimal, innovatively delineated various technology-related indicators from the National Research and Education Network technology domain as distinct model technology-related constructs, for example, the measurements of core network traffic, while also integrating indicators for context-related constructs, such as the spectrum of institutions typically serviced by National Research and Education Networks. Employing secondary data from the Trans-European Research and Education Network Association's annual Compendiums, the study undertook a Partial Least Squares regression analysis to empirically evaluate this autoregressive model instantiation to ascertain key model parameters, such as indicator loadings and path coefficients. The study also engaged in an extensive reliability and validity analysis of this model instantiation, affirming the empirical analysis's repeatability and the model instantiation's internal consistency in providing technology forecasting outputs in the National Research and Education Network technology domain.
Next, the study developed a cross-sectional model instantiation for the National Research and Education Network technology domain. Although incapable of longitudinal technology forecasting, this model instantiation marked a considerable improvement in performance over the autoregressive model instantiation. Its development not only amalgamated knowledge from action research conducted within the South African National Research Network and insights from the annual Compendiums of the Trans-European Research and Education Networking Association but also hypotheses from scholarly literature. Partial Least Squares regression analysis, employing data from the Trans-European Research and Education Network Association Compendiums, confirmed various hypothesised relationships, except the anticipated positive correlation between a National Research and Education Network's infrastructure and advanced services capabilities. This exception underscores the influence of technology leapfrogging within the National Research and Education Network community, potentially disrupting established technology development and adoption patterns.
The study concluded with an extensive analysis of the proposed framework's strengths and weaknesses. Informed by a broad scholarly discourse on the strengths and weaknesses of Data Fusion and Structural Equation Modelling, this evaluation scrutinised the framework's capability to incorporate context-related information in forecasting computations and its potential susceptibility to inaccuracies arising from structural model misspecification. This investigation employed various model instances specific to the National Research and Education Network technology domain for this assessment, including a structurally disarranged model instantiation.
This study heralds a transformative advancement in Technology Forecasting, particularly within the National Research and Education Networks technology domain, by introducing a novel amalgamation of Structural Equation Modelling and context-sensitive Data Fusion to perform transversal and longitudinal technology prediction using technology and context-related indicators. This innovative approach contributes to Engineering and Technology Management by offering an advanced tool for strategic planning and technology trend analysis. The core publications from the study demonstrated the development, practical application and assessment of this integrated framework. In contrast, the study’s supplementary publications enhanced the understanding and application of Partial Least Squares analysis tools to perform the regression analysis required to create various model instantiations for the National Research and Education Networks technology domain.