reccomendation_system.py

# -*- coding: utf-8 -*-
"""Reccomendation system.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1LcXsKpxEThboSebnaA45-Y-f7IKfujGK
"""

import pandas as pd
import numpy as np

df = pd.read_csv("/content/MOCK_DATA (1).csv")

df

import pandas as pd
from sklearn.metrics.pairwise import cosine_similarity

# Convert categorical variables into numerical representations
df['College_ID'] = df['College'].astype('category').cat.codes
df['Project_Domain_ID'] = df['Project Domain'].astype('category').cat.codes

# Create a user-item matrix
user_item_matrix = df.pivot_table(index='College_ID', columns='Project_Domain_ID', aggfunc=lambda x: 1, fill_value=0)

# Calculate cosine similarity
cosine_sim = cosine_similarity(user_item_matrix)

# Function to get project domain recommendations for a college
def get_project_domain_recommendations(college_id, user_item_matrix, cosine_sim):
    college_vector = user_item_matrix.loc[college_id].values.reshape(1, -1)
    sim_scores = cosine_similarity(college_vector, user_item_matrix)
    sim_scores = sim_scores.flatten()

    # Get indices of project domains with the highest similarity scores
    similar_project_domains = sim_scores.argsort()[::-1][1:]

    # Get recommendations based on similar project domains
    recommendations = user_item_matrix.columns[similar_project_domains]

    return recommendations

# Example: Get recommendations for College ID 0 (JOHNS)
college_id = 1
project_domain_recommendations = get_project_domain_recommendations(college_id, user_item_matrix, cosine_sim)

print(f"Recommendations for College ID {college_id}:\n{project_domain_recommendations}")

from sklearn.model_selection import train_test_split
from sklearn.metrics import precision_score, recall_score

df['College_ID'] = df['College'].astype('category').cat.codes
df['Project_Domain_ID'] = df['Project Domain'].astype('category').cat.codes

# Create a user-item matrix
user_item_matrix = df.pivot_table(index='College_ID', columns='Project_Domain_ID', aggfunc=lambda x: 1, fill_value=0)

# Split the data into training and testing sets
train_data, test_data = train_test_split(df[['College_ID', 'Project_Domain_ID']], test_size=0.2, random_state=42)

def get_project_domain_recommendations(college_id, user_item_matrix, cosine_sim):
    college_vector = user_item_matrix.loc[college_id].values.reshape(1, -1)
    sim_scores = cosine_similarity(college_vector, user_item_matrix)
    sim_scores = sim_scores.flatten()

    # Get indices of project domains with the highest similarity scores
    similar_project_domains = sim_scores.argsort()[::-1][1:]

    # Get recommendations based on similar project domains
    recommendations = user_item_matrix.columns[similar_project_domains]

    return recommendations

# Evaluate precision and recall on the test set
precision_values = []
recall_values = []

for college_id in test_data['College_ID'].unique():
    # Get actual liked project domains for the college in the test set
    actual_liked_project_domains = set(test_data[test_data['College_ID'] == college_id]['Project_Domain_ID'])

    # Get recommended project domains
    recommended_project_domains = set(get_project_domain_recommendations(college_id, user_item_matrix, cosine_sim))

    # Calculate precision and recall
    precision = len(actual_liked_project_domains.intersection(recommended_project_domains)) / len(recommended_project_domains)
    recall = len(actual_liked_project_domains.intersection(recommended_project_domains)) / len(actual_liked_project_domains)

    precision_values.append(precision)
    recall_values.append(recall)