Table 3 Open in new tab Performance...

Species

Method^a

PRE

F-value

ACC

MCC

E. coli

1. RF

0.828

0.820

0.830

0.824

0.825

0.650

2. SVM

0.798

0.790

0.800

0.794

0.795

0.590

3. LightGBM

0.806

0.830

0.800

0.818

0.815

0.630

4. KNN

0.862

0.750

0.880

0.802

0.815

0.635

5. LR

0.814

0.790

0.820

0.802

0.805

0.610

{1, 2}

0.842

0.800

0.850

0.821

0.825

0.651

{1, 2, 3}

0.830

0.830

0.830

0.660

{1, 2, 3, 4}

0.845

0.820

0.850

0.832

0.835

0.670

{1, 2, 3, 4, 5}

0.840

0.790

0.850

0.814

0.820

0.641

{1, 3, 4}*

0.856

0.830

0.860

0.843

0.845

0.690

M. musculus

1. RF

0.810

0.843

0.804

0.826

0.823

0.647

2. SVM

0.818

0.829

0.817

0.824

0.823

0.647

3. LightGBM

0.810

0.826

0.807

0.818

0.817

0.633

4. KNN

0.810

0.729

0.831

0.768

0.780

0.563

5. LR

0.808

0.829

0.804

0.818

0.817

0.634

{1, 2}

0.821

0.826

0.821

0.823

0.647

{1, 2, 3}

0.807

0.823

0.804

0.815

0.813

0.627

{1, 2, 3, 4}

0.826

0.839

0.824

0.833

0.832

0.663

{1, 2, 3, 4, 5}

0.828

0.836

0.827

0.832

0.663

{1, 2, 4}*

0.835

0.829

0.837

0.832

0.833

0.667

H. sapiens

1. RF

0.834

0.843

0.834

0.839

0.838

0.677

2. SVM

0.837

0.839

0.837

0.838

0.677

3. LightGBM

0.854

0.863

0.854

0.859

0.858

0.717

4. KNN

0.833

0.749

0.850

0.789

0.800

0.603

5. LR

0.840

0.823

0.844

0.831

0.833

0.667

{1, 2}

0.855

0.846

0.857

0.850

0.852

0.703

{1, 2, 3}

0.855

0.846

0.857

0.850

0.852

0.703

{1, 2, 3, 4}

0.863

0.846

0.867

0.855

0.857

0.713

{1, 2, 3, 4, 5}

0.863

0.846

0.867

0.855

0.857

0.713

{3, 4, 5}*

0.867

0.849

0.870

0.858

0.860

0.720

Table 3

Open in new tab

Performance comparison of different single method-based models and a selection of ensemble models for predicting malonylation sites of the three species on the independent test

Species	Method^a	PRE	SN	SP	F-value	ACC	MCC
E. coli	1. RF	0.828	0.820	0.830	0.824	0.825	0.650
2. SVM	0.798	0.790	0.800	0.794	0.795	0.590
3. LightGBM	0.806	0.830	0.800	0.818	0.815	0.630
4. KNN	0.862	0.750	0.880	0.802	0.815	0.635
5. LR	0.814	0.790	0.820	0.802	0.805	0.610
{1, 2}	0.842	0.800	0.850	0.821	0.825	0.651
{1, 2, 3}	0.830	0.830	0.830	0.830	0.830	0.660
{1, 2, 3, 4}	0.845	0.820	0.850	0.832	0.835	0.670
{1, 2, 3, 4, 5}	0.840	0.790	0.850	0.814	0.820	0.641
{1, 3, 4}*	0.856	0.830	0.860	0.843	0.845	0.690
M. musculus	1. RF	0.810	0.843	0.804	0.826	0.823	0.647
2. SVM	0.818	0.829	0.817	0.824	0.823	0.647
3. LightGBM	0.810	0.826	0.807	0.818	0.817	0.633
4. KNN	0.810	0.729	0.831	0.768	0.780	0.563
5. LR	0.808	0.829	0.804	0.818	0.817	0.634
{1, 2}	0.821	0.826	0.821	0.823	0.823	0.647
{1, 2, 3}	0.807	0.823	0.804	0.815	0.813	0.627
{1, 2, 3, 4}	0.826	0.839	0.824	0.833	0.832	0.663
{1, 2, 3, 4, 5}	0.828	0.836	0.827	0.832	0.832	0.663
{1, 2, 4}*	0.835	0.829	0.837	0.832	0.833	0.667
H. sapiens	1. RF	0.834	0.843	0.834	0.839	0.838	0.677
2. SVM	0.837	0.839	0.837	0.838	0.838	0.677
3. LightGBM	0.854	0.863	0.854	0.859	0.858	0.717
4. KNN	0.833	0.749	0.850	0.789	0.800	0.603
5. LR	0.840	0.823	0.844	0.831	0.833	0.667
{1, 2}	0.855	0.846	0.857	0.850	0.852	0.703
{1, 2, 3}	0.855	0.846	0.857	0.850	0.852	0.703
{1, 2, 3, 4}	0.863	0.846	0.867	0.855	0.857	0.713
{1, 2, 3, 4, 5}	0.863	0.846	0.867	0.855	0.857	0.713
{3, 4, 5}*	0.867	0.849	0.870	0.858	0.860	0.720

Species	Method^a	PRE	SN	SP	F-value	ACC	MCC
E. coli	1. RF	0.828	0.820	0.830	0.824	0.825	0.650
2. SVM	0.798	0.790	0.800	0.794	0.795	0.590
3. LightGBM	0.806	0.830	0.800	0.818	0.815	0.630
4. KNN	0.862	0.750	0.880	0.802	0.815	0.635
5. LR	0.814	0.790	0.820	0.802	0.805	0.610
{1, 2}	0.842	0.800	0.850	0.821	0.825	0.651
{1, 2, 3}	0.830	0.830	0.830	0.830	0.830	0.660
{1, 2, 3, 4}	0.845	0.820	0.850	0.832	0.835	0.670
{1, 2, 3, 4, 5}	0.840	0.790	0.850	0.814	0.820	0.641
{1, 3, 4}*	0.856	0.830	0.860	0.843	0.845	0.690
M. musculus	1. RF	0.810	0.843	0.804	0.826	0.823	0.647
2. SVM	0.818	0.829	0.817	0.824	0.823	0.647
3. LightGBM	0.810	0.826	0.807	0.818	0.817	0.633
4. KNN	0.810	0.729	0.831	0.768	0.780	0.563
5. LR	0.808	0.829	0.804	0.818	0.817	0.634
{1, 2}	0.821	0.826	0.821	0.823	0.823	0.647
{1, 2, 3}	0.807	0.823	0.804	0.815	0.813	0.627
{1, 2, 3, 4}	0.826	0.839	0.824	0.833	0.832	0.663
{1, 2, 3, 4, 5}	0.828	0.836	0.827	0.832	0.832	0.663
{1, 2, 4}*	0.835	0.829	0.837	0.832	0.833	0.667
H. sapiens	1. RF	0.834	0.843	0.834	0.839	0.838	0.677
2. SVM	0.837	0.839	0.837	0.838	0.838	0.677
3. LightGBM	0.854	0.863	0.854	0.859	0.858	0.717
4. KNN	0.833	0.749	0.850	0.789	0.800	0.603
5. LR	0.840	0.823	0.844	0.831	0.833	0.667
{1, 2}	0.855	0.846	0.857	0.850	0.852	0.703
{1, 2, 3}	0.855	0.846	0.857	0.850	0.852	0.703
{1, 2, 3, 4}	0.863	0.846	0.867	0.855	0.857	0.713
{1, 2, 3, 4, 5}	0.863	0.846	0.867	0.855	0.857	0.713
{3, 4, 5}*	0.867	0.849	0.870	0.858	0.860	0.720

^aEach item in this column refers to a single method-based model or an ensemble model that was built based on combining different single models (e.g. ‘1. RF’ means the model is trained based on RF, while ‘{1, 2}’ stands for the ensemble model that is built based on combining the single models numbered ‘1’ and ‘2’).

^*The optimal ensemble model was selected by exhaustively examining all possible random combinations of up to five single models.

This Feature Is Available To Subscribers Only