In evaluating recommendation systems, it is essential to measure both how accurately the system predicts user preferences and how well it ranks recommended items. **Predictive accuracy metrics** such as mean squared error (MSE) quantify the average squared difference between predicted and actual ratings, providing a straightforward measure of model performance. **Ranking metrics**, on the other hand, assess how effectively the system orders recommendations so that users are likely to find the most relevant items near the top of their recommendation lists. Together, these metrics offer a comprehensive view of a recommendation engine's strengths and weaknesses.


import unittest
import user_code
import ast
import re   
import importlib
import csv
import unittest
import user_code
import importlib
import numpy as np

class TestTask(unittest.TestCase):

    def test_precision_at_3(self):
        import user_code
        importlib.reload(user_code)
        predicted = [2.5, 4.0, 3.5, 5.0, 1.0]
        actual = [3.0, 5.0, 2.0, 4.0, 1.0]
        result = user_code.evaluate_predictions(predicted, actual)
        _dynamic_test(
            self,
            result is not None and abs(result[1] - 2/3) < 1e-6,
            "Correct precision@3 for provided example.",
            f"Expected precision@3 0.666..., got {result[1] if result else result}",
        )

    def test_short_lists(self):
        import user_code
        importlib.reload(user_code)
        predicted = [3.0, 4.0]
        actual = [4.0, 2.0]
        result = user_code.evaluate_predictions(predicted, actual)
        # MSE = ((3-4)^2 + (4-2)^2)/2 = (1+4)/2 = 2.5
        # Precision@3: only 2 items, so overlap of top-2/3
        pred_indices = np.argsort(predicted)[::-1][:3]
        actual_indices = np.argsort(actual)[::-1][:3]
        overlap = len(set(pred_indices) & set(actual_indices))
        expected_precision = overlap/3
        _dynamic_test(
            self,
            result is not None and abs(result[0] - 2.5) < 1e-6,
            "Correct MSE for short lists.",
            f"Expected MSE 2.5, got {result[0] if result else result}",
        )
        _dynamic_test(
            self,
            abs(result[1] - expected_precision) < 1e-6,
            "Correct precision@3 for short lists.",
            f"Expected precision@3 {expected_precision}, got {result[1] if result else result}",
        )

def _dynamic_test(test_case, condition, success_message, failure_message):
    if condition:
        test_case._testMethodName = success_message
        test_case.assertTrue(True, success_message)
    else:
        test_case._testMethodName = failure_message
        test_case.fail(failure_message)

def normalize_text(text):
    text = text.lower()
    text = re.sub(r"\\s{2,}", " ", text)
    text = re.sub(r"\\s*([,:?])\\s*", r"\\1 ", text)
    return text.strip()

def change_var(code: str, var_name: str, value: str) -> str:
    tree = ast.parse(code)
    lines = code.splitlines()
    changed = False
    # Collect all assignment nodes to modify
    assign_nodes = [
        (i, node)
        for i, node in enumerate(tree.body)
        if isinstance(node, ast.Assign)
        and any(isinstance(target, ast.Name) and target.id == var_name for target in node.targets)
    ]

    # If nothing to change, return unmodified code
    if not assign_nodes:
        return code

    # Perform replacements for all matching assignments (from last to first to not break line offsets)
    for i, node in reversed(assign_nodes):
        start_line = node.lineno - 1
        line = lines[start_line]
        indent = ' ' * (len(line) - len(line.lstrip()))
        lines[start_line] = f"{indent}{var_name} = {value}"
        next_line = len(lines)
        for next_node in tree.body[i+1:]:
            if hasattr(next_node, 'lineno'):
                next_line = next_node.lineno - 1
                break
        if next_line > start_line + 1:
            lines[start_line+1:next_line] = []
        changed = True

    return '\\n'.join(lines) if changed else code

if __name__ == "__main__":
    unittest.main()


test_main.py

Master the skills needed to design, implement, and scale modern recommendation engines and transactional association models. Go from analyzing raw retail checkouts using classic algorithmic principles like Apriori and FP-Growth to building advanced personalization systems via collaborative filtering and matrix factorization. Learn how to engineer data pipelines that power modern digital storefront discovery layers and drive user engagement metrics.

Discover how to find hidden relationships within massive transactional data structures. Learn the core mathematical frameworks retail giants use to optimize shelf placement, inventory structures, and product bundling strategies.

What is a latent feature in the context of recommendation systems?

How does SVD help in predicting missing ratings?

What does a low mean squared error indicate?

Which metric is best for evaluating ranking quality?

What is the main challenge of matrix factorization on sparse data?

How can dimensionality reduction improve recommendation quality?

Take your algorithmic knowledge to the next level by handling heavy corporate transaction data without memory crashes. Master high-performance compressed trees and discover how to eliminate commercial noise from rule patterns.

Get hands-on experience building the recommendation engines that power modern digital platforms. Learn how to look past product groupings to map historical user behaviors, ratings, and shared consumer habits.

Dive into the mathematical fundamentals of advanced machine learning models for sparse datasets. Learn how to uncover latent features hidden inside multi-dimensional user feedback data to accurately predict missing interactions.

Challenge: Calculating Predictive Accuracy and Ranking Metrics

Solution