Skip to main content

Table 11 Correlations between used features and not used non-indicator features for module 3

From: Sparsifying machine learning models identify stable subsets of predictive features for behavioral detection of autism

Feature

A2

A4

A8

B3_miss

B2

B7

B8

D4

D3_miss

Male

B10

0.641

0.270

0.422

− 0.07

0.425

0.172

0.554

0.228

− 0.07

0.023

B11

0.487

0.195

0.317

− 0.06

0.324

0.171

0.521

0.362

− 0.06

0.061

A1

0.387

0.293

0.328

− 0.03

0.189

0.163

0.343

0.037

− 0.00

− 0.03

Age_months

0.308

0.119

0.085

0.006

0.257

− 0.01

0.334

0.095

− 0.00

0.024

A3

0.335

0.252

0.167

− 0.09

0.321

0.087

0.361

0.273

− 0.09

0.048

A5

0.231

0.285

0.173

− 0.07

0.256

0.071

0.341

0.294

− 0.08

0.023

A7

0.366

0.086

0.274

− 0.05

0.222

0.122

0.320

0.052

− 0.06

0.025

A6

0.481

0.346

0.385

− 0.12

0.338

0.165

0.495

0.246

− 0.14

0.054

C2

0.365

0.184

0.408

− 0.06

0.250

0.157

0.388

0.168

− 0.05

0.126

C1

0.347

0.147

0.370

− 0.03

0.271

0.125

0.369

0.127

− 0.04

0.099

E1

0.133

0.076

0.171

− 0.03

0.103

0.133

0.149

0.155

− 0.03

0.059

E3

0.029

0.100

0.075

− 0.02

0.101

0.033

0.062

0.068

− 0.02

0.001

E2

0.254

0.078

0.230

− 0.02

0.122

0.108

0.218

0.243

− 0.03

0.031

B9

0.455

0.271

0.329

− 0.09

0.381

0.188

0.518

0.337

− 0.08

0.041

B4

0.180

0.113

0.138

− 0.02

0.116

0.170

0.170

0.139

− 0.02

0.011

B5

0.476

0.147

0.255

− 0.07

0.345

0.105

0.406

0.079

− 0.06

0.033

B6

0.369

0.101

0.267

− 0.07

0.237

0.087

0.287

0.125

− 0.06

0.023

B1

0.261

0.141

0.141

− 0.11

0.283

0.064

0.278

0.221

− 0.09

0.057

B3

0.507

0.171

0.320

− 0.04

0.480

0.090

0.502

0.184

− 0.04

0.017

D2

0.059

0.148

0.075

− 0.04

0.108

0.092

0.087

0.180

− 0.03

0.026

D3

0.105

0.051

0.117

− 0.00

0.044

− 0.00

0.140

0.073

− 0.00

0.039

D1

0.261

0.216

0.171

− 0.06

0.215

0.020

0.255

0.312

− 0.04

0.116

  1. This shows the high correlations between features and so, the need for a robust feature selection technique and classification