alt="Schematic illustration of fuzzy c-means cluster analysis of Influenza A (H1N1) infected human."/>
Figure 2.10 Fuzzy c-means clustering algorithm of Influenza A (H1N1) infected human.
Figure 2.11 Fuzzy c-means clustering algorithm of Influenza A (H1N1) infected human.
After that, type-I and type-II errors are used for finding the accuracy and predicting the output between the actual and predicted values. Table 2.2 is represented as rows and columns where rows are experiment and columns are represented as gold set. The meaning of true positive (TP) is that the set of glycans are identified by our experiment as same as the result mentioned in the gold set. True negative (TN) means that the set of glycans are identified by our experiment not same as the result mentioned in the gold set. False negative (FN) means that the set of glycans are identified by our experiment which is missing in our experiment. False positive (FP) means that the set of glycans are identified by our experiment that positive but missing in the gold set. Type-I and Type-II errors are described in Figure 2.12.
Table 2.1 Significant glycan list.
| Sr. no. | Structure |
|---|---|
| 1 | Neu5Aca2-3(6-O-Su)Galb1-4(Fuca1-3)GlcNAcb-Sp8 |
| 2 | Neu5Aca2-6Galb1-4GlcNAcb1-3Galb1-4(Fuca1-3)GlcNAcb1-3Galb1-4(Fuca1-3) GlcNAcb-Sp0 |
| 3 | Galb1-4GlcNAcb1-2Mana1-3(Neu5Aca2-6Galb1-4GlcNAcb1-2Mana1-6) Manb1-4GlcNAcb1-4GlcNAcb-Sp12 |
| 4 | GlcAb1-3GlcNAcb-Sp8 |
| 5 | Mana1-2Mana1-2Mana1-3(Mana1-2Mana1-6(Mana1-2Mana1-3)Mana1-6)Mana-Sp9 |
| 6 | GlcNAcb1-2Mana1-3(Galb1-4GlcNAcb1-2Mana1-6) Manb1-4GlcNAcb1-4GlcNAc-Sp12 |
| 7 | Galb1-4GlcNacb1-2(Galb1-4GlcNacb1-4)Mana1-3(Galb1-4GlcNacb1-2(Galb1-4GlcNacb1-6)Mana1-6)Manb1-4GlcNacb1-4GlcNacb-Sp21 |
| 8 | Galb1-3Galb1-4GlcNAcb-Sp8 |
| 9 | Galb1-3(Neu5Aca2-6)GalNAca-Sp14 |
| 10 | Neu5Aca2-6Galb1-4Glcb-Sp0 |
| 11 | Neu5Aca2-3Galb1-4GlcNAcb1-2Mana1-3(Neu5Aca2-3Galb1-4GlcNAcb1-2Mana1-6)Manb1-4GlcNAcb1-4GlcNAcb-Sp12 |
| 12 | Neu5Aca2-6Galb1-4GlcNAcb1-2Mana1-3(Neu5Aca2-3Galb1-4GlcNAcb1-2Mana1-6)Manb1-4GlcNAcb1-4GlcNAcb-Sp12 |
| 13 | Neu5Aca2-6GlcNAcb1-4GlcNAcb1-4GlcNAc-Sp21 |
| 14 | Neu5Aca2-3Galb1-4GlcNAcb-Sp8 |
| 15 | Neu5Aca2-6Galb1-4GlcNAcb-Sp0 |
| 16 | Neu5Gca2-3Galb1-4(Fuca1-3)GlcNAcb-Sp0 |
| 17 | Neu5Aca2-3Galb1-4GlcNAcb1-3Galb1-3GlcNAcb-Sp0 |
| 18 | Neu5Aca2-6Galb1-4GlcNAcb1-3Galb1-6GlcNAcb-Sp8 |
| 19 | Neu5Aca2-3Galb1-3GalNAcb1-4(Neu5Aca2-8Neu5Aca2-3)Galb1-4Glcb-Sp0 |
Table 2.2 The tabular format has been created from the above diagram.
| Our experiment | Gold Set | ||
|---|---|---|---|
| Positive (+) | Negative (−) | ||
| Positive (+) | T_Pos | F_Pos | |
| Negative (−) | F_Neg | T_Neg |
The performance of the method has been validated using various statistical measurements & metrices. For details, please refer to Table 2.3. The visual representation of the method's performance has also been depicted in Figure 2.13.
In our experiment, the value of t-pos = 14, F_pos = 5, F_Neg = 5, and T_neg = 418. Performance of our method using various metrices.
Figure 2.12 Concepts of type-I and type-II error in terms set.
Table 2.3 Performance of the method using various metrices.
| Parameter | Value |
|---|---|
| Sensitivity | 0.736 |
| True negative rate | 0.976 |
| Precision | 0.736 |
| Negative-predictive value | 0.163 |
| Miss rate | 0.263 |
| Fall out | 0.011 |
| False discovery rate | 0.357 |
| False omission rate | 0.011 |
| Threat score | 0.583 |
| Prevalence threshold | 0.149 |
| Accuracy | 0.977 |
| Balance accuracy | 0.856 |
| Matthews correlation coefficient | 0.725 |
| Fowlkes-Mallows index | 0.736 |
|
|
