Variation-Oriented Data Filtering for Improvement in Model Complexity of Air Pollutant Prediction Model

Algorithm 1

int 𝑖=1;       //initialize the beginning of the segment

int 𝑗=2;       //initialize the end of the segment

select (𝐱₁, 𝑦₁);      //select the data at the first day of the first segment

for 𝑗=2 to 𝑁 do   //loop from 2 to 𝑁, i.e., the end of the series

   if 𝑖=𝑗−1 and (2) is true {

      //if the 𝑖th day and the 𝑗th day are successive days and the variation of

      the SO₂ levels in these two days is equal to or greater than 𝛿

    select (𝐱_𝑗, 𝑦_𝑗)     //select the data from the 𝑗th day

    𝑖=𝑗;       //change the end of the segment to the beginning of the segment

    𝑗++;       //change the end of the segment to the next day

}

elseif (2) is true {//if the 𝑖th day and the 𝑗th day are not on successive days but

       their variation is equal to or greater than 𝛿

   select (𝐱_𝑗−1, 𝑦_𝑗−1)  //select data at the (𝑗−1)th day

   select (𝐱_𝑗, 𝑦_𝑗)      //select data at the 𝑗th day

   𝑖=𝑗;       //set the beginning of the next segment

   𝑗++; }      //change the end of the segment to the next day

else 𝑗++; }       //discard the data point and move to next segment