Marshhhhh / first

0 stars 0 forks source link

1 #14

Open Marshhhhh opened 7 years ago

Marshhhhh commented 7 years ago

Взаимодействие номинативных предикторов при построении модели с параметрическими переменными зачастую сами переменные используются в исходном виде. Однако часто существуют строгие различия по влиянию совокупности значений переменных на целевую функцию. Так, например, по матрице можем увидеть различие по переменным mosaicplot() Дополнение такого различия позволяет улучшить модель

http://blog.alphaarchitect.com/2015/05/05/value-investing-research-simple-methods-to-improve-the-piotroski-f-score/#gs.WkSaIp0

http://www.statsoft.com/textbook/credit-scoring

http://www.sciencedirect.com/science/article/pii/S1877050913002494

http://www.sciencedirect.com/science/article/pii/S1879933714000050

http://users.uoa.gr/~mpournaras/credit-risk-modeling---part-3.html

https://www.google.com/patents/US20060212386

https://cran.r-project.org/doc/contrib/Sharma-CreditScoring.pdf http://www.algorithmist.ru/2012/05/random-forest-r.html http://r-analytics.blogspot.ru/2015/04/blog-post.html#.WE2GSNWLTIX

Marshhhhh commented 7 years ago

http://trevorstephens.com/kaggle-titanic-tutorial/r-part-5-random-forests/