The 3-Step Method
Step 1: Verify all input intervals are equal (same ฮx).
Step 2: Compute AROC = ฮy/ฮx for each consecutive interval.
Step 3: Check differences between consecutive AROCs.
โข Differences = 0 โ Linear
โข Differences are constant (non-zero) โ Quadratic
โข Differences not constant โ Neither
| x | f(x) |
|---|---|
| 1 | 4 |
| 2 | 7 |
| 4 | 10 |
| 8 | 13 |
| Interval | AROC = ฮy/ฮx |
|---|---|
| [1, 2] | (7โ4)/(2โ1) = 3 |
| [2, 4] | (10โ7)/(4โ2) = 3/2 |
| [4, 8] | (13โ10)/(8โ4) = 3/4 |
Careful: unequal intervals here
In Example 1, the x-intervals are NOT equal (ฮx = 1, 2, 4). This means we can't use the "AROC pattern" test for linear vs quadratic โ we can only verify constancy of slope. Since the slopes differ, it's not linear. The linear/quadratic identification method requires equal ฮx.
| x | g(x) |
|---|---|
| โ3 | 9 |
| โ1 | 1 |
| 1 | 1 |
| 3 | 9 |
| 5 | 25 |
| Interval (ฮx=2) | AROC | ฮ between AROCs |
|---|---|---|
| [โ3, โ1] | (1โ9)/2 = โ4 | โ |
| [โ1, 1] | (1โ1)/2 = 0 | +4 |
| [1, 3] | (9โ1)/2 = 4 | +4 |
| [3, 5] | (25โ9)/2 = 8 | +4 |
| x | f(x) | AROC |
|---|---|---|
| 1 | 0 | โ |
| 2 | 1 | 1/1 = 1 |
| 3 | 4 | 3/1 = 3 |
| 4 | 9 | 5/1 = 5 |
| x | g(x) | AROC |
|---|---|---|
| 1 | 0 | โ |
| 2 | 1 | 1/1 = 1 |
| 5 | 4 | 3/3 = 1 |
| 10 | 9 | 5/5 = 1 |
| x | h(x) | AROC (ฮx=2) |
|---|---|---|
| 1 | โ1 | โ |
| 3 | 1 | 2/2 = 1 |
| 5 | 2 | 1/2 = 1/2 |
| 7 | 2 | 0/2 = 0 |
Connection between concavity and function type
Linear: AROC is constant โ neither concave up nor concave down.
Quadratic (concave up): AROC is increasing โ positive leading coefficient (parabola opens up).
Quadratic (concave down): AROC is decreasing โ negative leading coefficient (parabola opens down).
| x | k(x) | AROC (ฮx=0.1) |
|---|---|---|
| 1 | 4 | โ |
| 1.1 | 1 | โ3/0.1 = โ30 |
| 1.2 | โ1 | โ2/0.1 = โ20 |
| 1.3 | โ2 | โ1/0.1 = โ10 |
| x | m(x) | AROC (ฮx=0.1) |
|---|---|---|
| 1 | 1 | โ |
| 1.1 | 4 | 3/0.1 = 30 |
| 1.2 | 7 | 3/0.1 = 30 |
| 1.3 | 10 | 3/0.1 = 30 |
| x | p(x) | AROC (ฮx=0.1) |
|---|---|---|
| 1 | 1 | โ |
| 1.1 | 7 | 6/0.1 = 60 |
| 1.2 | 11 | 4/0.1 = 40 |
| 1.3 | 13 | 2/0.1 = 20 |
Concave up even when decreasing!
In Example 4a, k(x) is decreasing (negative AROCs) yet concave up (AROCs increasing from โ30 to โ10). This is the key distinction from Topic 1.1 โ a function can be decreasing AND concave up at the same time. Concavity is about whether the slope is getting bigger or smaller, not about whether the function itself is going up or down.