Topsis-swastik-102316020 1.0.2

TOPSIS Method for Multi-Criteria Decision Making

Topsis-Swastik-102316020

TOPSIS — Technique for Order of Preference by Similarity to Ideal Solution
A Multi-Criteria Decision Making (MCDM) method for ranking alternatives.

---

Installation

pip install Topsis-swastik-102316020

---

Command Line Usage

topsis <InputFile> <Weights> <Impacts> <OutputFile>

Example

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

---

Python API Usage

from topsis_pkg import run_topsis

run_topsis(
    input_file  = "data.csv",
    weights     = "1,1,1,2",
    impacts     = "+,+,-,+",
    output_file = "result.csv"
)

---

Input File Format

Your CSV must have:

• Column 1: Alternative names (text)
• Columns 2 to last: Numeric criteria values

Example data.csv

Fund	Return	Risk	Expense	Liquidity
M1	0.94	0.88	6.3	40.0
M2	0.76	0.58	7.0	45.2
M3	0.93	0.86	3.4	60.6

---

Parameters

Weights

Comma-separated positive numbers — one per criterion column.
"1,1,1,2" means the 4th criterion is twice as important.

Impacts

Comma-separated + or - — one per criterion column.

Symbol	Meaning	Example
+	Higher value is better (benefit)	Return, Score, Quality
-	Lower value is better (cost)	Expense, Risk, Price

---

Output

The result CSV contains all original columns plus two new ones:

Column	Description
Topsis Score	Score from 0 to 1. Higher = closer to ideal best
Rank	1 = best alternative

Example Output

Fund	Return	Risk	Expense	Liquidity	Topsis Score	Rank
M1	0.94	0.88	6.3	40.0	0.7542	1
M3	0.93	0.86	3.4	60.6	0.6218	2
M2	0.76	0.58	7.0	45.2	0.3782	3

---

Validations

The package checks for:

• Correct number of command-line parameters
• File existence (FileNotFoundError with clear message)
• Minimum 3 columns in input file
• Numeric values in criteria columns
• Matching count of weights, impacts, and columns
• Valid impact values (+ or - only)
• Comma-separated format for weights and impacts

---

Algorithm Steps

1. Normalize the decision matrix (Euclidean norm)
2. Weight the normalized matrix
3. Find Ideal Best (V⁺) and Ideal Worst (V⁻)
4. Calculate separation distances (d⁺ and d⁻)
5. Compute performance score: P = d⁻ / (d⁺ + d⁻)
6. Rank by descending score

---

License

MIT License — see LICENSE

---

Author

Swastik — [e-mail]
GitHub · PyPI