Topsis-Anushka-102003207 1.3.1

A python package to implement TOPSIS on a given dataset

TOPSIS-Python

Submitted By: Anushka Gargya
Roll Number: 102003207
Batch: 3CS11

---

pypi: https://pypi.org/project/Topsis-Anushka-102003207
git: https://github.com/agargya24/topsis-anushka

---

Installation

pip install Topsis-Anushka-102003207

What is TOPSIS

Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) originated in the 1980s as a multi-criteria decision making method. TOPSIS chooses the alternative of shortest Euclidean distance from the ideal solution, and greatest distance from the negative-ideal solution. More details at wikipedia.

How to use this package:

Topsis-Anushka-102003207 can be run as in the following example:

In Command Prompt

>> topsis data.csv "1,1,1,1" "+,+,-,+" result.csv

Sample dataset

Consider this sample.csv file

First column of file is removed by model before processing so follow the following format.

All other columns of file should not contain any categorical values.

Model	P1	P2	P3	P4	P5
M1	0.85	0.72	4.6	41.5	11.92
M2	0.66	0.44	6.6	49.4	14.28
M3	0.9	0.81	6.7	66.5	18.73
M4	0.8	0.64	6.9	69.7	19.51
M5	0.84	0.71	4.7	36.5	10.69
M6	0.91	0.83	3.6	42.3	11.91
M7	0.65	0.42	6.9	38.1	11.52
M8	0.71	0.5	3.5	60.9	16.4

weights vector = [ 1,2,1,2,1 ]

impacts vector = [ +,-,+,+,- ]

input:

topsis sample.csv "1,2,1,2,1" "+,-,+,+,-" output.csv

output:

output.csv file will contain following data :

Model	P1	P2	P3	P4	P5	Topsis score	Rank
M1	0.85	0.72	4.6	41.5	11.92	0.3267076760116426	6
M2	0.66	0.44	6.6	49.4	14.28	0.6230956090525585	2
M3	0.9	0.81	6.7	66.5	18.73	0.5006083702087599	5
M4	0.8	0.64	6.9	69.7	19.51	0.6275096427934269	1
M5	0.84	0.71	4.7	36.5	10.69	0.3249142875298663	7
M6	0.91	0.83	3.6	42.3	11.91	0.2715902624653612	8
M7	0.65	0.42	6.9	38.1	11.52	0.5439263412940541	4
M8	0.71	0.5	3.5	60.9	16.4	0.6166791918077927	3

The decision matrix (a) should be constructed with each row representing a Model alternative, and each column representing a criterion like Accuracy, R², Root Mean Squared Error, Correlation, and many more.

Weights (w) is not already normalised will be normalised later in the code.

Information of benefit positive(+) or negative(-) impact criteria should be provided in I.

The rankings are stored in a csv file, with the 1st rank offering us the best decision, and last rank offering the worst decision making, according to TOPSIS method.

Debugging and Exception Handling

The program has several assert statements which raise errors with helpful description in the following cases:

• Wrong dimensions of decision matrix (not 2D), weights (not 1D)
• Length of weights and impacts don't match
• Weights or impacts don't match number of attributes
• For command line, number of arguments is less than 3 required
• File extension must be .csv

License

Copyright 2023 Anushka Gargya
This repository is licensed under the MIT license.
See LICENSE for details. MIT