Table 2 Ranked performance of unsupervised dimension reduction techniques on the real datasets.

From: Nonlinear machine learning pattern recognition and bacteria-metabolite multilayer network analysis of perturbed gastric microbiome

Method

Paroni Sterbini

Amir3

Amir4

mean

PSI-ROC

 HD

2

2

1

1.67

 MCE

1

4

2

2.33

 MDSwUF

5

1

3

3.00

 PCA

3

3

4

3.33

 NMDS

3

6

5

4.67

 MDStyc

5

4

5

4.67

 MDSbc

7

6

5

6.00

PSI-PR

 HD

2

2

1

1.67

 MCE

1

5

2

2.67

 MDSwUF

5

1

3

3.00

 PCA

3

3

5

3.67

 MDStyc

5

3

5

4.33

 MDSbc

7

5

3

5.00

 NMDS

4

7

7

6.00

  1. The table shows the ranked performance of unsupervised dimension reduction techniques according to the PSI indices for sample separation (PSI-ROC and PSI-PR) in the space of the first two dimensions of embedding, for the three studied datasets (Paroni Sterbini, Amir3 and Amir4). Each rank is related to the results obtained in Table 1. The results are ordered by the mean performance (fourth column) from the best (top) to the worst (bottom) method.
  2. HD high dimension, MCE Minimum Curvilinear Embedding, MDSbc Multidimensional Scaling with Bray-Curtis dissimilarity, MDSwUF Multidimensional Scaling with weighted UniFrac distance, NMDS Nonmetric Multidimensional Scaling, MDStyc Multidimensional Scaling with Theta-YC distance, PCA Principal Component Analysis, PSI-ROC Projection Separability Index measured by Area Under the Curve, PSI-PR Projection Separability Index measured by Area Under the Precision Recall.
Back to article page