|
Full Spectrum Matching Method
This algorithm takes each peak in the unknown and attempts to find a corresponding matching peak in the library entry being searched. For a Forward Peak search, the X value of the first unknown peak is subtracted from the X value of each peak in the library entry being searched and the minimum absolute distance is saved. A score for this unknown peak is assigned based on its X distance according to the following table. Note that the X distance is calculated based on distance in data points not distance in X units.
Absolute Peak X
Distance |
Score for Peak |
| 0 or 1 |
100 |
| 2 |
80 |
| 3 |
40 |
| 4 |
20 |
| 5 |
10 |
| 6 |
8 |
| 7 |
4 |
| 8 |
2 |
| 9 |
1 |
| greater than 10 |
0 |
| Peak Search X Position Hit Scores |
Next, the hit amplitude of the unknown peak (values 1 to 9) is subtracted from the value of the nearest library peak, the absolute value times 10 is subtracted from the hit value for the peak determined above and the result is added to a total score for the current library entry. If the result is less than 0, no hit value is added to the total. This process is repeated for each peak in the unknown, then the total is normalized by dividing by the number of peaks in the unknown to get a hit score between 0 and 100.
| Absolute Peak Amplitude Difference |
Subtracted from Peak Score |
| 0 |
0 |
| 1 |
10 |
| 2 |
20 |
| 3 |
30 |
| 4 |
40 |
| 5 |
50 |
| 6 |
60 |
| 7 |
70 |
| 8 |
80 |
| Peak Search Y Value Hit Scores |
Note that the Forward peak search algorithm penalizes a library entry for not having a peak near an unknown peak, but does not penalize the library entry for having more peaks than the unknown.
In a Reverse peak search, each peak in a library peak table is compared to the peaks in the unknown and a score calculated according to the weighting described above. In this case, the normalization factor is the number of peaks in the current library entry to get a value between 0 and 100%. The Reverse peak search does not penalize the unknown for having more peaks than the library entry, but it may result in strong matches for library entries which have only a few peaks but which are well correlated with the corresponding peaks in the unknown. | 
0 Items