Solution 1 :------

I'll solve this step by step by computing the perceptron output for each input pair and comparing it
to the expected AND gate behavior.

Given:

• Weights: w₁ = 0.3, w₂ = 0.3

• Bias: b = -0.4

• Activation function: output = 1 if weighted sum ≥ 0, otherwise 0

Perceptron formula: Weighted sum = w₁x₁ + w₂x₂ + b = 0.3x₁ + 0.3x₂ - 0.4

Let me compute each case:

Case 1: (x₁, x₂) = (0, 0)

• Weighted sum = 0.3(0) + 0.3(0) - 0.4 = -0.4

• Since -0.4 < 0 → Output = 0

• Expected AND output = 0

• ✓ Correct

Case 2: (x₁, x₂) = (0, 1)

• Weighted sum = 0.3(0) + 0.3(1) - 0.4 = 0.3 - 0.4 = -0.1

• Since -0.1 < 0 → Output = 0

• Expected AND output = 0

• ✓ Correct

Case 3: (x₁, x₂) = (1, 0)

• Weighted sum = 0.3(1) + 0.3(0) - 0.4 = 0.3 - 0.4 = -0.1

• Since -0.1 < 0 → Output = 0

• Expected AND output = 0

• ✓ Correct

Case 4: (x₁, x₂) = (1, 1)

• Weighted sum = 0.3(1) + 0.3(1) - 0.4 = 0.3 + 0.3 - 0.4 = 0.2

• Since 0.2 ≥ 0 → Output = 1

• Expected AND output = 1

• ✓ Correct
Conclusion: Yes, the perceptron correctly mimics the AND gate behavior for all input combinations.
The weights w₁ = 0.3, w₂ = 0.3, and bias b = -0.4 are properly configured to implement the logical
AND function.

The key insight is that the bias of -0.4 ensures that both inputs must be 1 (contributing 0.6 total) to
overcome the negative bias and produce a positive weighted sum.

Solution 2:--------

I'll perform a forward pass through this neural network step by step.

Given Network Architecture:

• Input layer: 2 neurons

• Hidden layer: 2 neurons (ReLU activation)

• Output layer: 1 neuron (linear activation)

Given Parameters:

• Input-to-hidden weights: w₁,₁=0.1, w₁,₂=0.2, w₂,₁=0.3, w₂,₂=0.4

• Hidden layer biases: b₁=0.1, b₂=0.1

• Hidden-to-output weights: w₃,₁=0.5, w₃,₂=0.6

• Output bias: b₃=0.2

• Input vector: x=[1, 2]

Step 1: Input Layer to Hidden Layer

For hidden neuron 1:

• z₁ = w₁,₁x₁ + w₂,₁x₂ + b₁

• z₁ = 0.1(1) + 0.3(2) + 0.1

• z₁ = 0.1 + 0.6 + 0.1 = 0.8

For hidden neuron 2:

• z₂ = w₁,₂x₁ + w₂,₂x₂ + b₂

• z₂ = 0.2(1) + 0.4(2) + 0.1

• z₂ = 0.2 + 0.8 + 0.1 = 1.1

Step 2: Apply ReLU Activation to Hidden Layer

ReLU(z) = max(0, z)

• h₁ = ReLU(z₁) = ReLU(0.8) = 0.8

• h₂ = ReLU(z₂) = ReLU(1.1) = 1.1

Step 3: Hidden Layer to Output Layer

• z₃ = w₃,₁h₁ + w₃,₂h₂ + b₃

• z₃ = 0.5(0.8) + 0.6(1.1) + 0.2

• z₃ = 0.4 + 0.66 + 0.2 = 1.26

Step 4: Apply Linear Activation to Output

Since the output uses linear activation:

• y = z₃ = 1.26

Final Answer: The final output of the neural network is 1.26.

Summary of Forward Pass:

• Input: [1, 2]

• Hidden layer pre-activation: [0.8, 1.1]

• Hidden layer post-activation (ReLU): [0.8, 1.1]

• Output layer pre-activation: 1.26

• Final output (linear): 1.26

Solution 4:------

K-Means Clustering Analysis (k=3, 2 iterations)

Given Data Points

Point Coordinates

A1 (2,10)

A2 (2,6)

A3 (11,11)

A4 (6,9)

A5 (6,4)

A6 (1,2)

A7 (5,10)

A8 (4,9)

A9 (10,12)

A10 (7,5)
Point Coordinates

A11 (9,11)

A12 (4,6)

A13 (3,10)

A14 (3,8)

A15 (6,11)

Note: A2 appears twice in the original data - I'll use both instances as (2,6) and (4,6)

Step 1: Choose Initial Centroids (Random Selection)

Let's randomly select 3 initial centroids:

• C1 = A2 = (2,6)

• C2 = A9 = (10,12)

• C3 = A10 = (7,5)

ITERATION 1

Step 2: Calculate Euclidean Distances and Assign Clusters

For each point, calculate distance to each centroid using: d = √[(x₂-x₁)² + (y₂-y₁)²]

Point Coords Dist to C1(2,6) Dist to C2(10,12) Dist to C3(7,5) Closest Cluster

A1 (2,10) 4.00 8.25 7.07 C1 1

A2 (2,6) 0.00 10.77 5.10 C1 1

A3 (11,11) 10.05 1.41 7.21 C2 2

A4 (6,9) 5.00 5.00 4.12 C3 3

A5 (6,4) 4.47 10.77 1.41 C3 3

A6 (1,2) 4.12 14.87 6.71 C1 1

A7 (5,10) 4.24 5.39 5.39 C1 1

A8 (4,9) 3.61 7.62 4.47 C1 1

A9 (10,12) 10.77 0.00 7.62 C2 2

A10 (7,5) 5.10 7.62 0.00 C3 3

A11 (9,11) 8.06 1.41 6.32 C2 2

A12 (4,6) 2.00 8.49 3.16 C1 1

A13 (3,10) 4.12 7.28 5.83 C1 1

A14 (3,8) 2.24 8.54 4.47 C1 1

A15 (6,11) 5.66 4.12 6.08 C2 2

Step 3: Update Centroids (End of Iteration 1)

Cluster 1: A1(2,10), A2(2,6), A6(1,2), A7(5,10), A8(4,9), A12(4,6), A13(3,10), A14(3,8)

• New C1 = ((2+2+1+5+4+4+3+3)/8, (10+6+2+10+9+6+10+8)/8) = (3.0, 7.625)

Cluster 2: A3(11,11), A9(10,12), A11(9,11), A15(6,11)

• New C2 = ((11+10+9+6)/4, (11+12+11+11)/4) = (9.0, 11.25)

Cluster 3: A4(6,9), A5(6,4), A10(7,5)

• New C3 = ((6+6+7)/3, (9+4+5)/3) = (6.33, 6.0)

ITERATION 2

Step 4: Recalculate Distances with New Centroids

Point Coords Dist to C1(3.0,7.625) Dist to C2(9.0,11.25) Dist to C3(6.33,6.0) Closest Cluster

A1 (2,10) 2.45 7.18 5.72 C1 1

A2 (2,6) 1.73 9.49 4.33 C1 1

A3 (11,11) 8.06 2.06 6.40 C2 2

A4 (6,9) 3.54 3.81 3.06 C3 3

A5 (6,4) 4.61 8.50 2.01 C3 3

A6 (1,2) 5.66 12.79 6.25 C1 1

A7 (5,10) 2.83 4.61 4.12 C1 1

A8 (4,9) 1.56 5.59 3.61 C1 1

A9 (10,12) 7.28 1.25 6.71 C2 2

A10 (7,5) 4.33 6.71 1.15 C3 3

A11 (9,11) 6.32 0.25 5.39 C2 2

A12 (4,6) 1.73 7.35 2.33 C1 1

A13 (3,10) 2.38 6.02 5.00 C1 1

A14 (3,8) 0.62 6.52 3.54 C1 1

A15 (6,11) 4.24 3.04 5.10 C2 2

Step 5: Final Cluster Assignments (After Iteration 2)

Cluster 1 (Red): A1, A2, A6, A7, A8, A12, A13, A14

• Points: (2,10), (2,6), (1,2), (5,10), (4,9), (4,6), (3,10), (3,8)

Cluster 2 (Blue): A3, A9, A11, A15

• Points: (11,11), (10,12), (9,11), (6,11)

Cluster 3 (Green): A4, A5, A10

• Points: (6,9), (6,4), (7,5)

Final Centroids (After Iteration 2)

• C1 = (3.0, 7.625)

• C2 = (9.0, 11.25)

• C3 = (6.33, 6.0)

Summary

After 2 iterations of K-Means clustering with k=3:

1. Cluster 1 contains 8 points mainly in the left side of the coordinate space

2. Cluster 2 contains 4 points in the upper-right region

3. Cluster 3 contains 3 points in the middle-right region

The algorithm has converged to a stable partitioning where each cluster represents a distinct
spatial region of the data points.

Solution 5:-

I'll perform one step of forward propagation through this neural network with sigmoid activation.

Given:

• Input: X₁ = 1, X₂ = -1

• Hidden neuron: W₁ = 0.3, W₂ = -0.2, bias = 0.1

• Output neuron: W₃ = 0.4, bias = -0.1

• Target output: Y = 0.5

Step 1: Input to Hidden Layer

Calculate the weighted sum for the hidden neuron:

• z₁ = W₁X₁ + W₂X₂ + bias

• z₁ = 0.3(1) + (-0.2)(-1) + 0.1

• z₁ = 0.3 + 0.2 + 0.1 = 0.6

Apply sigmoid activation:

• h₁ = σ(z₁) = σ(0.6) = 1/(1 + e⁻⁰·⁶)

• h₁ = 1/(1 + 0.5488) = 1/1.5488 = 0.6457

Step 2: Hidden Layer to Output Layer

Calculate the weighted sum for the output neuron:

• z₂ = W₃h₁ + bias

• z₂ = 0.4(0.6457) + (-0.1)

• z₂ = 0.2583 - 0.1 = 0.1583

Apply sigmoid activation:

• ŷ = σ(z₂) = σ(0.1583) = 1/(1 + e⁻⁰·¹⁵⁸³)

• ŷ = 1/(1 + 0.8537) = 1/1.8537 = 0.5395

Final Results:

• Network output: ŷ = 0.5395

• Target output: Y = 0.5

• Error: |0.5395 - 0.5| = 0.0395

Summary of Forward Propagation:

• Input: [1, -1]

• Hidden layer pre-activation: z₁ = 0.6

• Hidden layer post-activation: h₁ = 0.6457

• Output layer pre-activation: z₂ = 0.1583

• Final output: ŷ = 0.5395

The network produces an output of 0.5395, which is very close to the target value of 0.5.

Solution 6:-

Simple Linear Regression Calculation

House Size (sq. ft.) Price ($1000s)

1400 245

1600 312

1700 279

1875 308

1100 199

Step 1: Calculate Basic Statistics

Sample size: n = 5

Calculate means:

• X̄ = (1400 + 1600 + 1700 + 1875 + 1100) ÷ 5 = 7675 ÷ 5 = 1535

• Ȳ = (245 + 312 + 279 + 308 + 199) ÷ 5 = 1343 ÷ 5 = 268.6

Step 2: Calculate Required Sums

X Y (X - X̄ ) (Y - Ȳ) (X - X̄ )(Y - Ȳ) (X - X̄ )²

1400 245 -135 -23.6 3,186 18,225

1600 312 65 43.4 2,821 4,225

1700 279 165 10.4 1,716 27,225

1875 308 340 39.4 13,396 115,600

1100 199 -435 -69.6 30,276 189,225

Sums:

• Σ(X - X̄ )(Y - Ȳ) = 3,186 + 2,821 + 1,716 + 13,396 + 30,276 = 51,395

• Σ(X - X̄ )² = 18,225 + 4,225 + 27,225 + 115,600 + 189,225 = 354,500

Step 3: Calculate Regression Coefficients

Slope (β₁): β₁ = Σ(X - X̄ )(Y - Ȳ) / Σ(X - X̄ )² β₁ = 51,395 / 354,500 = 0.145

Intercept (β₀): β₀ = Ȳ - β₁X̄ β₀ = 268.6 - (0.145 × 1535) β₀ = 268.6 - 222.575 = 46.025

Final Results

Regression Equation: Y = 46.025 + 0.145X

• β₀ (intercept) = 46.025 (in $1000s, so $46,025)

• β₁ (slope) = 0.145 (price increases by $145 per additional sq. ft.)

• The base price of a house (when size = 0) would be approximately $46,025

• For every additional square foot, the house price increases by approximately $145

• A 1500 sq. ft. house would be predicted to cost: 46.025 + 0.145(1500) = $263,525

Solution 7:-

Sales Regression Analysis

Given Data

Year (x) Sales (y)

2005 12

2006 19

2007 29

2008 37

2009 45

Step 1: Transform the Data

To simplify calculations, let's transform the years by subtracting 2005:

• x = 0 for 2005, x = 1 for 2006, x = 2 for 2007, x = 3 for 2008, x = 4 for 2009

xy

0 12

1 19

2 29

3 37

4 45

Step 2: Calculate Basic Statistics

Sample size: n = 5

Calculate means:

• x̄ = (0 + 1 + 2 + 3 + 4) ÷ 5 = 10 ÷ 5 = 2
 • ȳ = (12 + 19 + 29 + 37 + 45) ÷ 5 = 142 ÷ 5 = 28.4

Step 3: Calculate Required Sums

x y (x - x̄) (y - ȳ) (x - x̄)(y - ȳ) (x - x̄)²

0 12 -2 -16.4 32.8 4

1 19 -1 -9.4 9.4 1

2 29 0 0.6 0 0

3 37 1 8.6 8.6 1

4 45 2 16.6 33.2 4

Sums:

• Σ(x - x̄)(y - ȳ) = 32.8 + 9.4 + 0 + 8.6 + 33.2 = 84

• Σ(x - x̄)² = 4 + 1 + 0 + 1 + 4 = 10

Step 4: Calculate Regression Coefficients

Slope (b): b = Σ(x - x̄)(y - ȳ) / Σ(x - x̄)² b = 84 / 10 = 8.4

Intercept (a): a = ȳ - b·x̄ a = 28.4 - (8.4 × 2) a = 28.4 - 16.8 = 11.6

Step 5: Regression Equation (Transformed Scale)

y = 11.6 + 8.4x (where x = 0 corresponds to 2005)

Step 6: Convert Back to Original Scale

Since x = Year - 2005, we have: y = 11.6 + 8.4(Year - 2005) y = 11.6 + 8.4·Year - 8.4·2005 y = 11.6 +
8.4·Year - 16,842 y = -16,830.4 + 8.4·Year

Final Answer

A. Least Squares Regression Line: y = -16,830.4 + 8.4x

Where:

• y = sales (in units)

• x = year

• The slope (8.4) means sales increase by 8.4 units per year

• Sales are growing at a rate of 8.4 units annually

Verification

Let's verify with our data points:

• 2005: y = -16,830.4 + 8.4(2005) = -16,830.4 + 16,842 = 11.6 ≈ 12 ✓

• 2006: y = -16,830.4 + 8.4(2006) = -16,830.4 + 16,850.4 = 20 ≈ 19 ✓

• 2008: y = -16,830.4 + 8.4(2008) = -16,830.4 + 16,867.2 = 36.8 ≈ 37 ✓

• 2009: y = -16,830.4 + 8.4(2009) = -16,830.4 + 16,875.6 = 45.2 ≈ 45 ✓

Sales Estimates for Future Years

2010: y = -16,830.4 + 8.4(2010) = -16,830.4 + 16,884 = 53.6 units

2011: y = -16,830.4 + 8.4(2011) = -16,830.4 + 16,892.4 = 62.0 units

2012: y = -16,830.4 + 8.4(2012) = -16,830.4 + 16,900.8 = 70.4 units

I'll solve this KNN classification problem step by step.For the KNN classification of point P(5,3) with
K=3:

Step 1: Calculate Euclidean distances from P(5,3) to all points:

• d(P,A) = √[(5-2)² + (3-4)²] = √10 ≈ 3.16

• d(P,B) = √[(5-4)² + (3-2)²] = √2 ≈ 1.41

• d(P,C) = √[(5-4)² + (3-4)²] = √2 ≈ 1.41

• d(P,D) = √[(5-6)² + (3-2)²] = √2 ≈ 1.41

• d(P,E) = √[(5-6)² + (3-4)²] = √2 ≈ 1.41

• d(P,F) = √[(5-6)² + (3-6)²] = √10 ≈ 3.16

Step 2: Identify the 3 nearest neighbors: Four points (B, C, D, E) are tied with the minimum distance
of 1.41. Since we need K=3 neighbors, we select any 3 of these 4 points. Let's choose B, C, and D.

Step 3: Apply majority voting:

• B(4,2): Red class

• C(4,4): Blue class

• D(6,2): Blue class

Result: Point P(5,3) is classified as Blue since 2 out of 3 nearest neighbors belong to the Blue class.

The visualization above shows the dataset with the new point P and lines connecting it to its 3
nearest neighbors used for classification.