Skip to main content

Table 1 Feature descriptions

From: Explainable prediction of daily hospitalizations for cerebrovascular disease using stacked ensemble learning

Feature Name

Feature Descriptions

HAs Features

 HAs Lag x

Historical HAs on day x before the day for prediction, x ∈ {1, 2, 3, …, L}

 HAs Lag 1L mean

Moving average of historical HAs during the previous 1 to L days

 HAs Lag 1L std

Standard deviation of historical HAs during the previous 1 to L days

Environmental Features

 Pa Lag x

Historical values of P on day x before the day for prediction, x ∈ {1, 2, 3, …, L}

 P Lag 1L mean

Moving average of historical P during the previous 1 to L days

 P Lag 1L std

Standard deviation of historical P during the previous 1 to L days

Calendar Features

 DOW

Day of the week, {Mon., Tues., …, Sun.}—> {1, 2, …, 7}

 MON

Month of the year, {Jan., Feb., …, Dec.}—> {1, 2, …, 12}

 SEA

Season of the year, {spring, summer, fall, winter}—> {1, 2, 3, 4}

 YEAR

The year, {2015, 2016, 2017, 2018}—> {1, 2, 3, 4}

 TS

Timestamp, serial number from 1 to 1461

 HOL

Holiday, [0,1], 1 represented the day is a holiday, while 0 represented not

 WD

Workday, [0,1], 1 represented the day is a work day, while 0 represented not

 FWD

First work day, [0,1], 1 represented the day is the first workday, while 0 represented not

 LWD

Last work day, [0,1], 1 represented the day is the last work day, while 0 represented not

  1. aP ∈ {PM2.5, PM10, PMC, SO2, NO2, CO, O3, AQI, TEM, RH}