STA314: Assignment 1, Fall 2020
统计学代写 One (1) .R fifile called functions.R that contains all of the R functions you wrote.One (1) .R fifile called make_tidy.R that contains a script the···
Please submit your answers via Quercus as a single .zip fifile. This zip fifile should include: 统计学代写
One (1) .R fifile called functions.R that contains all of the R functions you wrote.
One (1) .R fifile called make_tidy.R that contains a script the defifines and registers the parsnip engine
for the functions defifined in functions.R.
A pdf of the compiled vignette.
The marking scheme is as follows:
40% for the functions in functions.R and the tidymodels wrapper defifined in make_tidy.R passing some unit tests. (See below for details)
20% for complete documentation (as described in the Week 4 tutorial) for all functions. This should include a simple use case for each function.
40% for the vignette. It should clearly step through the process of using your functions to complete the task using non-technical language where appropriate. All plots should have clear captions, axis names,and legends.
Please note: 统计学代写
The raw markdown fifile is not required.
Please note:
Please read the instructions very carefully. There is a restriction on the packages you are allowed to use in this assignment.
The task
Your task for this assignment is to program up functions to compute the logistic Lasso using coordinate descent on the the penalized iteratively reweighted least squares algorithm. Please note: You will need to modify the derivation of coordiante descent to allow for weights!
You will then build the helper functions required to add your algorihm as a parsnip model so that it can be used in a tidymodels workflflow.
You are to then to write a short vignette that explains how to use your function within a tidymodels workflflow to perform classifification. This should include how to tune the model. This vignette should not call any of the functions in the fifile functions.R directly, but should instead have the following code (or its equivalent) somewhere near the beginning
From that point forward, the vignette should use the functions you have written only through the tidymodels
interface you have built.
The functions 统计学代写
You only need to write two main function (although you’re welcome to use helper functions as long as they
are documented appropriately). It should have the following signature:
Unit tests 统计学代写
The exact unit tests that will be used for marking will not be released.
The purpose of these tests is to ensure that the functions work as indicated. Below is some skeleton code for a unit test for fit_logistic_lasso
The unit tests will test both functions and will test the tidymodels interface.
I strongly recommend that you write your own version of these tests to check your code. This is a situation where you can (and should!) collaborate with fellow students and compare your test code. As always, you are not allowed to collaborate on any other aspect of this assignment.
Which packages can you use
You can use any packages included when you run the following commands:
library(tidyverse)
library(tidymodels)
Note these are metapackages that include a whole host of other packages including dplyr, ggplot, and
parsnip among others.
In particular, you may not use any implementations of the logistic lasso in your submitted code (eg you can’t use the one in glmnet.). You may, however, fifind these implementations useful when you write your own tests. But please don’t submit this work. 统计学代写
更多其他:代写作业 数学代写 物理代写 生物学代写 程序编程代写 Reasoning and Learning