Build a KNN Classifier Practice Problem

This data science coding problem helps you practice K-Nearest Neighbors, build a knn classifier, and implementation skills. Read the problem statement, write your solution, and strengthen your understanding of K-Nearest Neighbors.

• Problem ID: 183
• Problem key: 183-build-a-knn-classifier
• URL: https://datacrack.app/solve/183-build-a-knn-classifier
• Difficulty: medium
• Topic: K-Nearest Neighbors
• Module: Supervised Learning

Problem Statement

# 🧩 Build a KNN Classifier

---

### 🎯 Goal

Combine distance calculation, neighbor selection, and majority voting into a complete KNN classifier.

---

### 📖 Introduction

You have now built each part of KNN separately. A full classifier repeats the same pipeline for every new query:

1. Calculate its distance to every training row.
2. Keep the `k` smallest distances.
3. Vote using their labels.
4. Return the winning label.

Unlike many machine-learning algorithms, KNN does not fit a formula during training. It stores the training examples and does this work when asked to predict.

---

### 💻 Task

Implement `knn_predict(X, y, queries, k)`.

Your function should:

- Predict a label for every query in `queries`.
- Use Euclidean distance.
- Select the `k` nearest training samples.
- Use majority voting.
- Return predictions in the same order as `queries`.
- Break vote ties alphabetically.

---

### 📥 Input / 📤 Output

**Input:** training features `X`, labels `y`, new samples `queries`, and `k`.

**Output:** a list of predicted labels.

---

### 🧩 Starter Code

```python
def knn_predict(X, y, queries, k):
    # X: training feature matrix
    # y: training labels
    # queries: new samples to predict
    # k: number of nearest neighbors

    def distance(row, target):
        # Calculate Euclidean distance
        pass

    def vote(labels):
        # Return the label with the most votes
        pass

    def predict_one(query):
        # Predict one query point using KNN
        pass

    # Predict all queries
    pass
```


---
### 💡 Example

**Input**

```python
knn_predict(
    X=[
        [1, 1],
        [2, 2],
        [5, 5],
        [6, 6]
    ],
    y=[
        "A",
        "A",
        "B",
        "B"
    ],
    queries=[
        [1.5, 1.5],
        [5.5, 5.5]
    ],
    k=3
)
```

**Expected output**

```python
[
    "A",
    "B"
]
```

---

### ⚠️ Common Mistakes

- Reusing the neighbors from the first query for later queries.
- Forgetting to return one result for every query.
- Voting before sorting by distance.

Starter Code

def knn_predict(X, y, queries, k):
    # X: training feature matrix
    # y: training labels
    # queries: new samples to predict
    # k: number of nearest neighbors

    def distance(row, target):
        # Calculate Euclidean distance
        pass

    def vote(labels):
        # Return the label with the most votes
        pass

    def predict_one(query):
        # Predict one query point using KNN
        pass

    # Predict all queries
    pass