| Abstrakt: |
Objective To analyze the epidemiological distribution characteristics of chickenpox in Urumqi City from 2015 to 2023, to construct a prediction model by combining Baidu search keywords, and to explore the complementary application of Baidu index in chickenpox prevention and monitoring. Methods Descriptive epidemiological methods were used to analyze the characteristics of varicella case triple distribution in Urumqi City from 2015 to 2023. Chickenpox keywords were identified and a comprehensive Baidu search index was constructed. The models ARIMA and ARIMAX were constructed, the prediction effectiveness of the two models was evaluated by mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE). Results The average annual reported incidence rate of varicella in Urumqi was 80.85/100 000, with a higher incidence in men than in women (χ² = 1.136, P = 0.287), there were statistically significant differences in incidence rates by age group (χ² = 7 582.372, P < 0.001), seven districts and one county had different average annual incidence rates (χ² = 21.496, P < 0.001), with the highest in the Toutunhe district (100.54/100 000). ARIMAX(1,1,0)(1,0,0)52 was selected as the best prediction model (prediction set MAE 12.04%, RMSE 13.80%, MASE 1.18%) with a good fitting effect. Conclusion The ARIMAX prediction model established based on the search term Baidu index has a certain degree of predictability and sensitivity, and can predict the epidemic trend of chickenpox in Urumqi in time, which can be used as a technical support and further expansion of the traditional monitoring and early warning system. [ABSTRACT FROM AUTHOR] |
