A Trend-Based Segmentation Method and the Support Vector Regression for Financial Time Series Forecasting

Algorithm 1

A pseudocode for TBSM in time series data.

Define: Threshold // cutting threshold

   X_Thld // horizontal area

   Y_Thld // vertical area

   X // a time series

   Y // stock price

1: Procedure TBSM(𝑇)

2: Let T be represented as 𝑋[1,2,…,𝑛], 𝑌[1,2,…,𝑛]

3:  n = 0

4: Draw a line between (𝑋₁,𝑌₁) and (𝑋_𝑛,𝑌_𝑛)

5: Max d = maximum distance of (𝑋_𝑖,𝑌_𝑖) to the line

6: If (Max 𝑑 > Threshold)

7:  Let (𝑋_𝑖,𝑌_𝑖) be the point with maximum distance

8:  For j = 𝑋₁:𝑋_𝑛

9:   If (|𝑋_𝑗−𝑋_𝑖| < X_Thld) and (|𝑌_𝑗−𝑌_𝑖| < Y_Thld)

10:    Then Point[𝑛]=[𝑋_𝑗,𝑌_𝑗], n = n + 1

11:   End If

12:  End For

13:  Select from Point[𝑛]:𝑋_𝑡1=Min(𝑋₀),𝑋_𝑡2=Max(𝑋_𝑛)

14:  Return: S1 = T [𝑋₁,𝑋_𝑡1]

15:      S2 = T [𝑋_𝑡2,𝑋_𝑛]

16: End If