FITTING TREND LINES AND FORECASTING
FITTING BY EYE: inadequate technique. We will look at the following two: 3-MEDIAN METHOD AND LEAST-SQUARES REGRESSION
Only works for the following original data: 1. When trend is clearly linear 2. Random 3. Secular (increasing or decreasing) Does not work on: 1. Cyclical trends 2. Seasonal trends
STEP 1: STATS - LIST 1: TIME - LIST 2: INDEPENDENT VARIABLE STEP 2: SET GRAPH - SETTING -TYPE: SCATTER - LIST 1: TIME - LIST 2: INDEPENDENT VARIABLE - SET STEP 3: CALC -LINEAR REG - LIST 1: TIME - LIST 2: INDEPENDENT VARIABLE - COPY FORMULA: y1 STEP 4: FORCASTING - y1 (x) -EXE
The y-intercept: when time code is at 0, please be careful and consider what the data is about. Eg: Our last example, the y- intercept is the day before the opening of the salon. This has no real meaning. The gradient (rate of change): very important. Eg: growing by approx 2 customers per day.
EXERCISE 4B pg 163 Q’s 1-10