门诊流感样病例人数预测模型数据

针对流感样病例监测，基于桐乡各医疗机构门诊就诊记录，获取全年门急诊就诊人数、流感样病例症状人数、年龄组及性别分布，进而预测和评价全市门诊流感样病例发病趋势，为流感防控工作及哨点医院监测工作提供新的思路和方法。1. 数据采集：从医疗机构取得匿名化样本统计数据：每日门急诊就诊人数、流感样病例症状人数、年龄组及性别分布；2. 数据预处理：对流感样病例数据进行预处理，以消除异常值和噪声，平滑数据，减少随机波等；3.特征工程：生成年龄组特征AGW与性别特征GW。AGW（t）=α1​×age_0_5(t)+α2​×age_5_15(t)+α3​×age_15_25(t)+α4​×age_25_60(t)+α5​×age_60_plus(t)，α1-α5为模型权重。GW（t）=γ1​×male_cases(t)+γ2×female_cases(t)。γ1,γ2为模型权重。4. 模型构建：构建预测模型，F(t) 为某天的流感样病例预测人数。F(t)= w1​×C(t−1)+w2​×C(t−2)+⋯+wn​×C(t−n) +β1*AGW（t）+β2*GW（t）。其中C代表每日流感样病例症状人数的取值函数；C(t) 是当天的流感样病例症状人数；C(t−n)为t−n日的流感样病例症状人数； wn为C(t−n)的权重； β1和β2是衍生特征的权重。

This dataset is designed for influenza-like illness (ILI) surveillance, built upon outpatient and emergency visit records from various medical institutions in Tongxiang. It collects annual statistics including total outpatient and emergency visits, number of ILI cases, as well as age-group and gender distributions, to predict and evaluate the incidence trend of outpatient ILI across the city, thereby providing innovative ideas and methods for influenza prevention and control and sentinel hospital surveillance work.
1. Data Collection: Anonymized aggregated statistical data is acquired from medical institutions, covering daily number of outpatient and emergency visits, number of ILI cases, and age-group and gender distributions.
2. Data Preprocessing: Preprocessing is conducted on the ILI-related data to eliminate outliers and noise, smooth the dataset, and reduce random fluctuations, etc.
3. Feature Engineering: Two derived features, AGW and GW, are generated. The formula for AGW(t) is: $AGW(t) = alpha_1 imes age_{0_5}(t) + alpha_2 imes age_{5_15}(t) + alpha_3 imes age_{15_25}(t) + alpha_4 imes age_{25_60}(t) + alpha_5 imes age_{60_plus}(t)$, where $alpha_1$ to $alpha_5$ are model weights. The formula for GW(t) is: $GW(t) = gamma_1 imes male_cases(t) + gamma_2 imes female_cases(t)$, where $gamma_1$ and $gamma_2$ are model weights.
4. Model Construction: A prediction model is built, where $F(t)$ represents the predicted number of ILI cases on a given day. The model formula is: $F(t) = w_1 imes C(t-1) + w_2 imes C(t-2) + dots + w_n imes C(t-n) + eta_1 imes AGW(t) + eta_2 imes GW(t)$. Here, $C$ denotes the value function of daily ILI case counts; $C(t)$ is the number of ILI cases on the current day; $C(t-n)$ is the number of ILI cases on day $t-n$; $w_n$ is the weight of $C(t-n)$; $eta_1$ and $eta_2$ are the weights of the derived features.

搜集汇总

数据集介绍

特点

该数据集为门诊流感样病例人数预测模型数据，包含624条记录，每日更新，用于预测流感样病例发病趋势，支持流感防控工作。