Browsing by Subject "Forecasting"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Open Access
    Effects of different missing data imputation techniques on the performance of undiagnosed diabetes risk prediction models in a mixed-ancestry population of South Africa
    (Public Library of Science, 2015) Masconi, Katya L; Matsha, Tandi E; Erasmus, Rajiv T; Kengne, Andre P
    BACKGROUND: Imputation techniques used to handle missing data are based on the principle of replacement. It is widely advocated that multiple imputation is superior to other imputation methods, however studies have suggested that simple methods for filling missing data can be just as accurate as complex methods. The objective of this study was to implement a number of simple and more complex imputation methods, and assess the effect of these techniques on the performance of undiagnosed diabetes risk prediction models during external validation. METHODS: Data from the Cape Town Bellville-South cohort served as the basis for this study. Imputation methods and models were identified via recent systematic reviews. Models’ discrimination was assessed and compared using C-statistic and non-parametric methods, before and after recalibration through simple intercept adjustment. RESULTS: The study sample consisted of 1256 individuals, of whom 173 were excluded due to previously diagnosed diabetes. Of the final 1083 individuals, 329 (30.4%) had missing data. Family history had the highest proportion of missing data (25%). Imputation of the outcome, undiagnosed diabetes, was highest in stochastic regression imputation (163 individuals). Overall, deletion resulted in the lowest model performances while simple imputation yielded the highest C-statistic for the Cambridge Diabetes Risk model, Kuwaiti Risk model, Omani Diabetes Risk model and Rotterdam Predictive model. Multiple imputation only yielded the highest C-statistic for the Rotterdam Predictive model, which were matched by simpler imputation methods. CONCLUSIONS: Deletion was confirmed as a poor technique for handling missing data. However, despite the emphasized disadvantages of simpler imputation methods, this study showed that implementing these methods results in similar predictive utility for undiagnosed diabetes when compared to multiple imputation.
  • Thumbnail Image
    Item
    Restricted
    Forecasting performance of an estimated DSGE model for the South African economy
    (Wiley, 2011) Alpanda, Sami; Kotze, Kevin; Woglom, Geoffrey
    We construct a small open-economy New Keynesian dynamic stochastic general equilibrium (DSGE) model for South Africa with nominal rigidities, incomplete international risk sharing and partial exchange rate pass-through. The parameters of the model are estimated using Bayesian methods, and its out-of-sample forecasting performance is compared with Bayesian vector autoregression (VAR), classical VAR and random-walk models. Our results indicate that the DSGE model generates forecasts that are competitive with those from other models, and it contributes statistically significant information to combined forecast measures.
  • Thumbnail Image
    Item
    Open Access
    Inhomogeneities in molecular layers of Mira atmospheres
    (2011) Wittkowski, M; Boboltz, D A; Ireland, M; Karovicova, I; Ohnaka, K; Scholz, M; van Wyk, F; Whitelock, P; Wood, P R; Zijlstra, A A
    Aims. We investigate the structure and shape of the photospheric and molecular layers of the atmospheres of four Mira variables.
  • Thumbnail Image
    Item
    Open Access
    Risk models to predict chronic kidney disease and its progression: a systematic review
    (Public Library of Science, 2012) Echouffo-Tcheugui, Justin B; Kengne, Andre P
    A systematic review of risk prediction models conducted by Justin Echouffo-Tcheugui and Andre Kengne examines the evidence base for prediction of chronic kidney disease risk and its progression, and suitability of such models for clinical use.
  • Thumbnail Image
    Item
    Open Access
    Risk models to predict hypertension: a systematic review
    (Public Library of Science, 2013) Echouffo-Tcheugui, Justin B; Batty, G David; Kivimäki, Mika; Kengne, Andre P
    BACKGROUND: As well as being a risk factor for cardiovascular disease, hypertension is also a health condition in its own right. Risk prediction models may be of value in identifying those individuals at risk of developing hypertension who are likely to benefit most from interventions. Methods and FINDINGS: To synthesize existing evidence on the performance of these models, we searched MEDLINE and EMBASE; examined bibliographies of retrieved articles; contacted experts in the field; and searched our own files. Dual review of identified studies was conducted. Included studies had to report on the development, validation, or impact analysis of a hypertension risk prediction model. For each publication, information was extracted on study design and characteristics, predictors, model discrimination, calibration and reclassification ability, validation and impact analysis. Eleven studies reporting on 15 different hypertension prediction risk models were identified. Age, sex, body mass index, diabetes status, and blood pressure variables were the most common predictor variables included in models. Most risk models had acceptable-to-good discriminatory ability (C-statistic>0.70) in the derivation sample. Calibration was less commonly assessed, but overall acceptable. Two hypertension risk models, the Framingham and Hopkins, have been externally validated, displaying acceptable-to-good discrimination, and C-statistic ranging from 0.71 to 0.81. Lack of individual-level data precluded analyses of the risk models in subgroups. CONCLUSIONS: The discrimination ability of existing hypertension risk prediction tools is acceptable, but the impact of using these tools on prescriptions and outcomes of hypertension prevention is unclear.