QQ plot of code validation

Overview

I will draw a plot to show the value of power under different value of type one control.

Setting:

row<-40 col<-100 num<-5 str<-0 times<-20 typeOne <- seq(0.05,0.95,0.05)

QQ plot of code validation

• type one error (reg, smo)

##  [1] 0.0595 0.1090 0.1555 0.1770 0.2755 0.3160 0.3495 0.4360 0.4505 0.4750
## [11] 0.5520 0.6120 0.6545 0.7100 0.7720 0.8140 0.8465 0.8845 0.9605

##  [1] 0.0590 0.1100 0.1585 0.1900 0.2675 0.3345 0.3460 0.3975 0.4620 0.5095
## [11] 0.5605 0.5945 0.6150 0.7085 0.7410 0.7920 0.8415 0.9080 0.9615

• permutation (reg, smo)

##  [1] 0.0542 0.1061 0.1613 0.2032 0.2530 0.3133 0.3707 0.4033 0.4632 0.4999
## [11] 0.5613 0.6034 0.6636 0.7021 0.7573 0.8002 0.8564 0.9050 0.9524

##  [1] 0.0531 0.1056 0.1514 0.1913 0.2544 0.3011 0.3425 0.3963 0.4554 0.5125
## [11] 0.5493 0.6008 0.6459 0.7030 0.7631 0.8179 0.8572 0.8988 0.9521

• plot

Result interpretation

We can see the power line of each method are all close to the true power line, which means our power calculation ways works well.

Question

Why professor metioned multi-regression and Lasso during the meeting?

Notes

• simulate with more data if possible
• cite the author
• understand the relation between methods and problems

future work

• find equitvalent function of ep