radiomics | python implementation of the Radiomics approach | Machine Learning library

My problem

I would like to build a logistic regression model with a high AUC in predicting a binary variable.

I would like to use the following approach (if feasible):

1. Use an elastic net model (glmnet) to reduce the predictors and find the best hyperparameter (alpha and lambda)

2. Combine the output of this model (a simple linear combination) with an additional predictor (the opinion of a super doctor superdoc) in a logistic regression model (=finalmodel), similar as described on page 26 in:

Afshar P, Mohammadi A, Plataniotis KN, Oikonomou A, Benali H. From Handcrafted to Deep-Learning-Based Cancer Radiomics: Challenges and Opportunities. IEEE Signal Process Mag 2019; 36: 132–60. Available here

Example data

As example data I have a dataset with many numeric predictors and a binary (pos/neg) outcome (diabetes).

I am trying to visualize features that were extracted from a dataset of CT-scans using radiomics. I extracted features using the Pyradiomics library, and now I have a few .csv files.

I found this R Library which can read a dataset and generate images for it. Reading the dataset works according to plan using the input = import_pyradiomics(dir = "/path/to/datadir") command, but when I try to visualize it using for instance plot_correlation_matrix(rdr = input, view_as = "heatmap"), one of the visualization functions, I get the following error:

Error: $ operator not defined for this S4 class

I read about S3 and S4 classes, and I think that [line 160] in the library is where this error occurs first as it attempts to call this '$' operator. I am currently using R version 4.0.2. It seems to me like this issue is caused by this R library relying on an older version of R, but it is relatively up to date. What can I do to resolve this issue? Do I have to go in and manually change the code in the library? Or is there an easier fix, such as casting the S4 class to an S3 class? I feel like I'm missing an easy solution because I'm not familiar with R.

I found a similar question here, but it doesn't answer my specific question. Answers on other questions regarding this topic also do not work for me unfortunately.

I'm trying to install a package from Anaconda Cloud called pyradiomics. So I wrote this Dockerfile:

I want to calculate ICC for multiple data.frame simultaneity. All my data.frame have the same structure (5 columns by 83 rows). I created a function (my_icc) that generated the ICC and put the result in a table. I saw in this web site that when you want to apply a function on multiple data.frame at the same time, you need to put your data into a list. (Apply a function to multiple dataframes) However, my function doesn’t seem to support list element, because I use the function select. I use select function because, I want to do the ICC with my 3:5 columns. If I apply my function on one data.frame at the time, I don’t have any problem.

I use pyinstaller to package my python program into executable program(exe).

my program version:

Pyinstaller :3.3.1

Python :3.6.1 Windows:10-10.0.16299-SP0

I can run the data_processing_gui.py file with the command "python data_processing_gui.py" in my command-line window without any ERROR.

I use the command to package my python program(which name is data_processing_gui.py):

I have installed a python library called PyRadiomics by doing: pip install pyradiomics

When I do pip freeze | grep pyra in the command prompt, I see pyradiomics==2.0.0 clearly showing that the library was installed.

When I start a python interpreter and do import pyradiomics, it does not work. I realized that I need to do import radiomics.

I just happened to figure this out by luck. How is someone supposed to know how to import a library into their python script after installing it using pip. It seems that currently, you could install a library with pip install some_name and have to do import some_other_name in your Python code. I have been using Python for a while but never came across this before which lead me to assume that some_name is always the same as some_other_name but I found out that this is not the case. How is someone supposed to know what that other name is after installing a library called some_name