Table 3.
Average Results for Each Signature Over 24 Classifiers Revealing 5-Protein Classifiers.
Overall (“AD” + “MCI”) | Test Set “AD” | Test Set “MCI” | ||||||
---|---|---|---|---|---|---|---|---|
Signature | Size | Overall | AD Er | NAD Er | AD Er | NAD Er | AD Er | NAD Er |
139 | 64 | 75 | 42 | 50 | 22 | 25 | ||
18-Protein signature | Error average | 24.34 | 9.02 | 15.33 | 3.66 | 4.79 | 5.36 | 10.53 |
Agr % | 82 | 86 | 80 | 91 | 90 | 76 | 58 | |
82 | 91 | 66 | ||||||
10-Protein signature | Error average | 25.98 | 7.83 | 18.15 | 2.45 | 7.64 | 5.38 | 10.52 |
Agr % | 81 | 88 | 76 | 94 | 85 | 76 | 58 | |
81 | 89 | 66 | ||||||
6-Protein signature | Error average | 24.44 | 8.60 | 15.84 | 2.02 | 5.70 | 6.58 | 10.14 |
Agr % | 82 | 87 | 79 | 95 | 89 | 70 | 59 | |
82 | 92 | 64 | ||||||
5-Protein signature | Error average | 23.20 | 7.71 | 15.49 | 1.78 | 4.75 | 5.93 | 10.73 |
Agr % | 83 | 88 | 79 | 96 | 90 | 73 | 57 | |
83 | 93 | 65 |
Abbreviations: AD, Alzheimer disease; MCI, mild cognitive impairment.
Note. This is a five-protein classifier where approximately ninety-seven percent efficacy was detected to predict AD using this data and each highlighted value showed the efficacy to predict AD in each class.