Rexis is an R package and Shiny application that serve as a holistic approach to the algorithm selection problem for machine learning classification problems. Rexis reads in a .csv data file with binary targets in the first column. Column headers are encouraged for ease of use. Rexis then predicts the best algorithms from its built in taxonomy of algorithms. Next, Rexis performs all recommended algorithms for the problem and reports their perforance. According to research performed at the Air Force Institute of Technology, Rexis recommends an excellent algorithm in 78% of test problems.
To get started, install the Rexis package to RStudio from the Rexis project:
devtools::install_github("marcchale/Rexis", INSTALL_opts=c("--no-multiarch"))
library(Rexis)
Now you’re ready to use Rexis. Run the following code and browse to the desired data set. The “Heart”heart.csv" data set is one example included in the “Rexis/data/” folder. Type ResultsVar into your consolde to see the performance of the analysis!
ResultsVar <- rexecute()
ResultsVar
If you prefer the Shiny interface, run the code below in the R console and select a .csv !!
run_my_app("RexisApp")
You can access additional help documentation for the rexicute and the rexicuteshiny functions in your R IDE.
? rexicute
? rexicuteshiny
R 3.6.3;
Python 3.7
Python Modules: Most users are able to use Rexis without implicitly installing specific Python Packages. However, if issues arise, ensure each of the following Python modules run properly on your system.
import pandas as pd
import numpy as np
import time
import os
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn import preprocessing
from sklearn.ensemble import RandomForestClassifier
from sklearn.naive_bayes import MultinomialNB
from sklearn.metrics import recall_score #,precision_score
from sklearn import svm
from sklearn.svm import SVR
from scipy.stats import t
from scipy.stats import spearmanr
from sklearn.tree import DecisionTreeClassifier
from ParamGet import ParamGet
from matplotlib.patches import Patch
from matplotlib.lines import Line2D
from pareto import identify_pareto
from adjustText import adjust_text
from matplotlib.text import OffsetFrom
If you would like to read the AFIT Thesis that used the Rexis software, please email marc.chale@afit.edu
If you are searching for information on Rexis, Pennsyvania, please see the following link Rexis, PA
If you are searching for information on Regis Philbin, please see the following link Regis