Test statistic field maxima

The basic procedure for RFT validations is outlined in Examples/Basic/Validating. This section presents code details for validating the probability with which test statistic field maxima survive particular thresholds.

See ./rft1d/examples/val_max_1_onesample_t_1d.py and other “val_max*.py” scripts

Validations for test statistic set-level and cluster-level inference are available in scripts titled ./rft1d/examples/val_upx*.py

Note

To see details of test statistic computation click on the “Source code” links above each figure.

Univariate

One-sample

(Source code, png, hires.png, pdf)

../../_images/val_max_1_onesample_t_1d.png

Two-sample

(Source code, png, hires.png, pdf)

../../_images/val_max_2_twosample_t_1d.png

Regression

(Source code, png, hires.png, pdf)

../../_images/val_max_3_regress_1d.png

One-way ANOVA

(Source code, png, hires.png, pdf)

../../_images/val_max_4_anova1_1d.png

Multivariate

One-sample

(Source code, png, hires.png, pdf)

../../_images/val_max_5_onesample_T2_1d.png

Note

Increasing the number of iterations will cause the results to converge to RFT expectations.

Two-sample

(Source code, png, hires.png, pdf)

../../_images/val_max_6_twosample_T2_1d.png

Canonical correlation analysis

(Source code, png, hires.png, pdf)

../../_images/val_max_7_cca_1d.png

MANOVA

(Source code, png, hires.png, pdf)

../../_images/val_max_8_manova1_1d.png