From: Performance of penalized maximum likelihood in estimation of genetic covariances matrices

Penalty | % reduction in average loss | % replicates with increased loss | ||||||||
---|---|---|---|---|---|---|---|---|---|---|

V∞ | V1 | CV3 | CV5 | L5% | V∞ | V1 | CV3 | CV5 | L5% | |

${\mathcal{P}}_{\lambda}^{\ell}$ | 71.3 | 60.6 | 55.9 | 50.4 | 68.3 | 7.3 | 8.7 | 14.6 | 14.6 | 12.0 |

${\mathcal{P}}_{\lambda}^{\ell 2}$ | 72.9 | 63.7 | 60.7 | 58.1 | 70.2 | 6.5 | 7.5 | 13.0 | 13.2 | 10.0 |

${\mathcal{P}}_{\beta}^{b}$ | 71.4 | 62.9 | 58.1 | 53.9 | 69.5 | 6.4 | 7.5 | 13.6 | 14.0 | 9.8 |

${\mathcal{P}}_{\mathbf{\Sigma}}$ | 70.6 | 60.6 | 54.9 | 52.7 | 64.1 | 4.6 | 8.9 | 15.4 | 15.5 | 15.6 |

${\mathcal{P}}_{\rho}$ | 72.0 | 62.9 | 54.4 | 51.6 | 70.5 | 4.0 | 7.1 | 9.9 | 10.2 | 9.2 |