Skip to main content

Table 1 Correlation coefficients between S&P 500 changes and the predictors with polynomial and linear regressions

From: Predicting stock market movements using network science: an information theoretic approach

Correlation matrix

Act. S&P

Act. S&P

Sqrs. S&P

Abs. S&P

Strength distribution

    

KLD 3

0.5628 / 0.6081

0.0895

0.6705

0.6454

KLD 6

0.5752 / 0.5984

0.0837

0.6823

0.6360

KLD 9

0.5333 / 0.5334

0.0582

0.6498

0.6725

KLD 13

0.4794 / 0.4886

0.0408

0.6182

0.6219

KLD All

0.5582 / 0.5635

0.1175

0.6521

0.6587

RS 3

0.4185 / 0.4455

0.0173

0.3811

0.6630

RS 6

0.4326 / 0.4845

0.0159

0.3838

0.6615

RS 9

0.4196 / 0.4855

0.0093

0.3750

0.6526

RS 13

0.4385 / 0.4674

0.0024

0.4077

0.6685

RS All

0.4065 / 0.4447

0.0134

0.3649

0.6552

Mean

0.4189 / 0.4583

0.0129

0.3640

0.6536

Variance

0.1487 / 0.1641

0.0175

0.3548

0.6407

Skewness

0.5471 / 0.5610

0.0644

0.6265

0.5716

Kurtosis

0.5425 / 0.5581

0.0192

0.4047

0.6532

Eigenvector centrality

    

Mean

0.1526 / 0.1795

0.0099

0.2591

0.5351

Median

0.3168 / 0.3200

0.0110

0.2720

0.5509

Maximum

0.4272 / 0.4494

0.0068

0.2175

0.4836

Betweenness centrality

    

Mean

0.0435 / 0.0482

0.0111

0.2797

0.5482

Median

0.0288 / 0.0289

0.0102

0.0089

0.0332

Maximum

0.0288 / 0.0288

0.0162

0.2445

0.4350

Network modularity

    

Modularity

0.2973 / 0.2982

0.0082

0.1503

0.3906

  1. Note. Correlations are significant at the 0.05 level
  2. Act.: Actual values of S&P 500 changes
  3. Sqrs.: Squared values of S&P 500 changes
  4. Abs.: Absolute values of S&P 500 changes
  5. *Polynomial regression second-order/third-order
  6. **Linear regression