2.1.3.4.2 Algorithm for Trend Analysis

  • Linear model
y_t=intercept + slope \times t
  • Quadratic model
y_t=intercept + B_1 \times t + B_2 \times t^2
  • Exponential growth model
y_t= B_0 \times B_1^t
  • Logistic model
y_t= \frac{1}{B_0+ B_1 \times B_2^t}
  • Statistics
Mean absolute percentage error: MAPE = \sum_{t=1}^{n}\frac{|(y_t - \hat{y}_t)/y_t|}{n}
Mean absolute deviation : MAD = \sum_{t=1}^{n}\frac{|y_t - \hat{y}_t|}{n}
Mean squared deviation : MSD = \sum_{t=1}^{n}\frac{|y_t - \hat{y}_t|^2}{n}
English | Deutsch | 日本語