Demo of pyhasse.fuzzy¶

Aim of the program¶

Application of fuzzy concepts within the framework of partial orders is one interesting option to take into regard that not any slight numerical difference is considered as relevant for a ranking.

If not already done¶

Comment out or delete the following line if the module is installed.

In [1]:

#!pip install pyhasse.fuzzy

import all necessary libraries¶

In [2]:

import pathlib
from pyhasse.core.csv_io import CSVReader
from pyhasse.core.order import Order
from pyhasse.fuzzy.calc import Fuzzy
import json
from IPython.core.display import display, HTML
from string import Template

In [3]:

TESTFILENAME = '/csvdata/a_test12.txt'
HERE = pathlib.Path('__file__').parent
csv = CSVReader(fn=str(HERE) + TESTFILENAME, ndec=3)
red = csv.calc_reduced_system()
order = Order(csv.dmred,
              csv.redrows,
              csv.cols)
zeta = order.calc_relatmatrix(csv.dmred,
                              csv.redrows,
                              csv.cols)
redrows = csv.redrows
down = order.calc_downset(zeta, redrows)
up = order.calc_upset(zeta, redrows)

Some basic data¶

In [4]:

# print(csv.rows)
# print(csv.cols)
# print(csv.prop)
# print(csv.obj)
# print(csv.objred)
# print(csv.data)

your/user input¶

In [5]:

ndec = 3 
acut = 0.6

In [6]:

rows = csv.rows
cols = csv.cols
dm = csv.data
obj = csv.obj
prop = csv.prop
fuzzy = Fuzzy(rows, cols, dm, obj, prop, ndec)
sh = fuzzy.calc_koskomatrix(dm, rows, cols)
relat = fuzzy.fuzzyproduct(sh)
lr = len(relat)
for i1 in range(0, lr):
    for i2 in range(0, lr):
        relat[i1][i2] = round(relat[i1][i2], ndec)
# Note, the values of relat matrix should be sorted increasingly
# and delivered to the user

crisp_relat = fuzzy.calc_crisp_matrix(relat, acut)
objred_fuzzy_act = fuzzy.generate_representative_elements(crisp_relat)

In [7]:

objred_fuzzy_act

Out[7]:

['0', '4', '12', '16']

In [8]: