endt 0.11.1

Enhanced Decision Tree

Enhanced Decision Tree (EDT)

What is Enhanced Decision Tree

TODO

Install

There is two option to install EDT library: from PyPi (easy) or from source.

PyPi

Use simple command:

pip install edt

That's it

From source

TODO

Get started

Here some simple example to use EDT

Classical Decision Tree

Construct classical Decision Tree on syntatic dataset:

import endt

x1 = endt.ContinuumData([1.,1.,2.,2.], name="x1")
x2 = endt.ContinuumData([100.,200.,100.,200.], name="x2")
y = endt.ContinuumData([10.,20.,30.,40])

tree = endt.Tree()
tree.fit([x1, x2], y)
print(tree)

res = tree.predict(1,2)
print("predict y(1,2) =", res)

Simple Enhanced Decision Tree

Add LSQ of order 2 in previous code:

import endt

x1 = endt.ContinuumData([1.,1.,2.,2.], name="x1")
x2 = endt.ContinuumData([100.,200.,100.,200.], name="x2")
y = endt.ContinuumData([10.,20.,30.,40])

funcs = endt.func.create_lsq_functions([x1, x2], order=2)

tree = endt.Tree(lsq_funcs=funcs)
tree.fit([x1, x2], y)
print(tree)

res = tree.predict(1,2)
print("predict y(1,2) =", res)

EDT from CSV

Let our dataset save is in the CSV file data.csv, containing two collumn of feature parameters and third collumn of reults

Create EDT from CSV file:

import endt
import numpy as np

data = np.loadtxt("data.csv").T
x = list(map(endt.ContinuumData, data[:-1]))
y = endt.ContinuumData(data[-1])
for i, xi in enumerate(x, start=1):
    xi.name = f"x{i}" # set name x1, x2, x3... for every feature parameters
funcs = endt.func.create_lsq_functions(x, order=2)
tree = endt.Tree(lsq_funcs=funcs)
tree.fit(x, y)
print(tree)

res = tree.predict(1,2)
print("predict y(1,2) =", res)

Full documentation

Read full documentation https://nanoworld_ml.gitlab.io/enhanceddecisiontree/