Model Forecast Results for Brucellosis Incidence in China, 2021

Yuqi Jiang's dataset contains forecast results for 2021 from a comparative study of seven time-series models predicting human brucellosis incidence in China. The models, including SARIMA, Holt-Winters, NNAR, ETS, TBATS, and Prophet, were trained on monthly and annual incidence rates from 2011 to 2020. The Holt-Winters multiplicative model was identified as the best-performing model for short-term forecasting.

Use Cases

• Comparing the performance of different time-series forecasting models based on metrics like RMSE, MAE, MAPE, and MASE.
• Short-term forecasting of brucellosis incidence in China based on identified seasonal patterns.
• Benchmarking new forecasting algorithms against established models like SARIMA and Prophet using the provided results.
• Informing public health surveillance and early warning systems based on the recommended Holt-Winters multiplicative model.

Strengths

• Directly compares seven distinct time-series models (SARIMA, Holt-Winters additive/multiplicative, NNAR, ETS, TBATS, Prophet).
• Model performance is evaluated with four specific error metrics: RMSE, MAE, MAPE, and MASE.
• Training data covers a 10-year period from January 2011 to December 2020.
• The dataset is licensed under CC-BY-4.0, allowing for reuse with attribution.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is very small at 5.5 KB, indicating it likely contains only summary forecast results, not the underlying raw incidence data.

Provenance

Source

Yuqi Jiang via figshare.

Collection Method

Results from a modeling study using monthly and annual brucellosis incidence rates as raw data, analyzed with R software.

Time Range

Forecast results are for 2021, based on training data from 2011-2020.

Freshness

Last updated 2026-03-25 17:39:55; freshness should be verified.

Geography

China.