Sharon Nwanamidwa

Dr. Sharon Nwanamidwa

orcid

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 September 2026 in Forecasting
Verified
Journal Article
Open access
PDF file
JTJohannes Tshepiso Tsoku
Johannes Tshepiso Tsoku
MSMalose Patrick Shogole
Malose Patrick Shogole
SNSharon Nwanamidwa
Sharon Nwanamidwa
LMLebotsa Daniel Metsileng
Lebotsa Daniel Metsileng
Forecasting equity returns remains challenging because financial markets exhibit nonlinear dynamics, volatility clustering, and complex temporal dependencies that are difficult to capture using a single modelling approach. Traditional statistical models can capture dependence and volatility dynamics, while deep learning models are capable of learning nonlinear temporal patterns. This study evaluates the forecasting performance of traditional statistical models (ARIMA and GARCH), deep learning models (TCN and GRU), and hybrid models (ARIMA-TCN, ARIMA-GRU, GARCH-TCN, and GARCH-GRU) for forecasting the FTSE/JSE Top 40 index in South Africa. Daily closing prices comprising 4159 observations from 2010 to 2026 were analysed, with model performance evaluated on log returns using MSE, RMSE, and MAE, with the Naïve model used as a benchmark. The results show that all competing models substantially outperformed the Naïve benchmark, while the TCN outperformed the standalone GRU. GARCH-based models also demonstrated strong forecasting performance, highlighting the relevance of volatility dynamics in forecasting equity returns. Although GARCH-GRU recorded the lowest MSE, RMSE, and MAE among the models evaluated, its performance was very similar to that of the GARCH (2,2) model. The Diebold-Mariano test further showed no statistically significant difference in predictive accuracy between GARCH-GRU and GARCH (2,2) (DM = 0.9495; p = 0.3424). These findings indicate that GARCH-GRU and GARCH (2,2) provide statistically comparable forecasting performance, suggesting that the additional complexity of the hybrid framework does not necessarily result in a significant improvement in predictive accuracy. This study contributes to the financial forecasting literature by providing empirical evidence from an emerging African equity market and demonstrating the importance of volatility modelling and complementary statistical–deep learning approaches for forecasting FTSE/JSE Top 40 log returns.

1 September 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
NMNtebogang Dinah Moroke
Ntebogang Dinah Moroke
SNSharon Nwanamidwa
Sharon Nwanamidwa
OMOlorato Semantha Makobea
Olorato Semantha Makobea
Typical machine learning forecasting approaches, optimized via mean squared error, struggle in decision settings requiring directional consistency. The Maximum Agreement Linear Predictor (MALP) framework addresses this by maximizing Lin’s Concordance Correlation Coefficient (CCC), but its meta-learner value under real, weak-signal conditions is untested. We evaluate a two-stage MALP-Meta architecture (four base models, a CCC-optimized meta-learner) on real hourly S&P 500 (SPY) data (n = 5032; 2023–2026), with every result cross-validated across two computing environments. Only Ridge achieves significant directional accuracy (52.7%, p = 0.037); no MALP-Meta variant improves on this, and we diagnose a previously undocumented pathology in which CCC optimization collapses toward a constant under weak signal. Of three temporal robustness strategies, only stable feature selection matches Ridge’s significance (52.5%, p = 0.048). Under an externally-defined stress regime (VIX>25), one meta-learner configuration improves simultaneously on concordance and direction, an unconfirmed signal (p = 0.291). Sensitivity analysis shows the primary result is fragile to the evaluation split. These findings affirm rather than overturn the conventional agreement–error trade-off: agreement-based meta-learning’s real-world value is modest, conditional on genuine market stress, and detectable only through rigorous cross-validation and sensitivity analysis.

Verified
Journal Article
Open access
PDF file
NMNtebogang Dinah Moroke
Ntebogang Dinah Moroke
SNSharon Nwanamidwa
Sharon Nwanamidwa
Volatility regime identification underpins risk management and portfolio allocation in quantitative finance, yet standard mixture models fail in heavy-tailed environments: components are consumed by outliers rather than genuine persistent regimes. We propose the Bayesian Spike-and-Slab Finite Mixture with Adaptive Tail Regularisation (BSS-FM-ATR), which resolves this at the component level via a spike-and-slab prior on the degrees-of-freedom parameter νk. A latent binary indicator assigns each component to a slab state (data-driven tail adaptation for genuine regimes) or a spike state (inert heavy-tail absorber for artefacts). Applied to 19 JSE blue-chip securities over December 2019 to December 2025—spanning the COVID-19 crash and Eskom load-shedding episodes—BSS-FM-ATR achieves the highest silhouette score (0.3809 on the full dataset), regime persistence (0.9391), and interpretability (0.800) across nine standard baselines, including Gaussian HMM, MS-AR, MS-GARCH(1,1), and Bayesian Changepoint detection, plus three outlier-component comparators (Contaminated–Normal Mixture, TCLUST, and robust Bayesian mixture). A complete rerun after removing the contaminated observations confirms that ARI (Δ=−0.0283) and persistence (Δ=+0.0105) remain stable; the silhouette reduction (from 0.3809 to 0.3773) is a positive finding: it confirms that the spike-state component R3 was correctly identified as a compact, well-separated artefact cluster whose removal reveals the genuine regime structure. This dual validation establishes BSS-FM-ATR’s role as both a regime identifier and a data quality filter. The method isolates a Yahoo Finance data-contamination artefact (n=21, January–February 2025) through a principled spike-and-slab mechanism, providing an explicit posterior probability p(γ3=0∣X)>0.99 of artefact status: a structural identifier that contaminated-normal and robust Bayesian alternatives cannot supply, establishing its value as both a regime identifier and a data quality filter for financial monitoring systems.