Topsis-Saksham-102303157 1.0.0

A Python package for TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) analysis

Topsis-Saksham-102303157

A Python package implementing TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution), a multi-criteria decision analysis method.

Description

TOPSIS is a method of compensatory aggregation that compares a set of alternatives by identifying weights for each criterion, normalizing scores for each criterion, and calculating the geometric distance between each alternative and the ideal alternative.

Installation

pip install Topsis-Saksham-102303157

Usage

Command Line

After installation, you can use the topsis command directly from the terminal:

topsis <InputDataFile> <Weights> <Impacts> <OutputResultFileName>

Example

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

Parameters

• InputDataFile: Path to the input CSV file
• Weights: Comma-separated weights for each criterion (e.g., "1,1,1,2")
• Impacts: Comma-separated impacts (+ or -) for each criterion (e.g., "+,+,-,+")
• OutputResultFileName: Path for the output CSV file

Input File Format

The input CSV file should have:

• First column: Names/identifiers of alternatives
• Remaining columns: Numeric values for each criterion

Example (data.csv):

Fund Name,P1,P2,P3,P4,P5
M1,0.67,0.45,6.5,42.6,12.56
M2,0.6,0.36,3.6,53.3,14.47
M3,0.82,0.67,3.8,63.1,17.1

Output

The program generates a CSV file with two additional columns:

• Topsis Score: The calculated TOPSIS score for each alternative
• Rank: Ranking based on TOPSIS score (1 = best)

Features

• Input validation (file existence, column count, numeric values)
• Proper error handling with descriptive messages
• Vector normalization method
• Weighted normalized decision matrix
• Euclidean distance calculation
• TOPSIS score and ranking

Requirements

• Python 3.6+
• pandas >= 1.0.0
• numpy >= 1.18.0

License

MIT License

Author

Saksham (Roll Number: 102303157)

Contributing

Contributions, issues, and feature requests are welcome!

Support

For support, email [e-mail]