9/30/24, 5:48 AM ChatGPT

[a] Predicting Time to Failure Using the Trained Linear Regression Model

Given:
Coefficients:
w0 = 50
​

w1 = −3
​

w2 = −0.5
​

w3 = −0.02
​

Features of the Machine:

Machine Age (x1 ) = 5 years
​

Operating Temperature (x2 ) = 60°C ​

Vibration Level (x3 ) = 10 mm/s

​

Linear Regression Model Equation:

Time to Failure = w0 + w1 x1 + w2 x2 + w3 x3 ​ ​ ​ ​ ​ ​ ​

Step-by-Step Calculation:
1. Compute the Contribution of Each Feature:
Machine Age Contribution:

w1 x1 = (−3) × 5 = −15
​ ​

Operating Temperature Contribution:

w2 x2 = (−0.5) × 60 = −30
​ ​

Vibration Level Contribution:

w3 x3 = (−0.02) × 10 = −0.2
​ ​

2. Sum All Contributions Along with the Intercept:

Time to Failure = 50 + (−15) + (−30) + (−0.2) = 50 − 15 − 30 − 0.2 = 4.8

Answer:

The predicted time to failure of the machine is 4.8 months.

[b] Calculating the Partial Derivative of the Cost Function with Respect to w1 ​

Given Dataset (Table II):

https://chatgpt.com/c/66f9eccd-6fe8-800e-95ae-f1014c8f2289 1/4
9/30/24, 5:48 AM ChatGPT

Machine Age (x1i )

​ Operating Temp (x2i ) ​ Vibration Level (x3i )​ Actual Time to Failure (yi )
​

1 5 60 10 110

2 8 75 15 80

3 3 50 8 130

Coefficients:
w0 = 50 ​

w1 = −3 ​

w2 = −0.5 ​

w3 = −0.02​

Cost Function J(w) Using Mean Squared Error (MSE):

N
1
J(w) = ∑ (yi − y^i )2
2N i=1
​ ​ ​ ​ ​

Where:
N = Number of instances = 3
y^i = Predicted time to failure for instance i
​ ​

Objective:

∂J
Find ∂w1 . ​
​

Step 1: Compute Predicted Values y^i for Each Instance ​ ​

For Machine 1:
Features:
x11 = 5
​

x21 = 60
​

x31 = 10
​

Predicted Time to Failure y^1 : ​ ​

y^1 = 50 + (−3)(5) + (−0.5)(60) + (−0.02)(10) = 50 − 15 − 30 − 0.2 = 4.8

​ ​

For Machine 2:
Features:
x12 = 8
​

x22 = 75
​

x32 = 15
​

https://chatgpt.com/c/66f9eccd-6fe8-800e-95ae-f1014c8f2289 2/4
9/30/24, 5:48 AM ChatGPT

Predicted Time to Failure y^2 : ​ ​

y^2 = 50 + (−3)(8) + (−0.5)(75) + (−0.02)(15) = 50 − 24 − 37.5 − 0.3 = −11.8

​ ​

For Machine 3:
Features:
x13 = 3 ​

x23 = 50 ​

x33 = 8 ​

Predicted Time to Failure y^3 : ​ ​

y^3 = 50 + (−3)(3) + (−0.5)(50) + (−0.02)(8) = 50 − 9 − 25 − 0.16 = 15.84

​ ​

Step 2: Compute the Error Term (yi ​ − y^i ) for Each Instance ​ ​

For Machine 1:

y1 − y^1 = 110 − 4.8 = 105.2

​ ​ ​

For Machine 2:

y2 − y^2 = 80 − (−11.8) = 80 + 11.8 = 91.8

​ ​ ​

For Machine 3:

y3 − y^3 = 130 − 15.84 = 114.16

​ ​ ​

Step 3: Compute (yi ​ − y^i ) × x1i for Each Instance

​ ​ ​

For Machine 1:

(y1 − y^1 ) × x11 = 105.2 × 5 = 526

​ ​ ​ ​

For Machine 2:

(y2 − y^2 ) × x12 = 91.8 × 8 = 734.4

​ ​ ​ ​

For Machine 3:

(y3 − y^3 ) × x13 = 114.16 × 3 = 342.48

​ ​ ​ ​

Step 4: Sum the Products

3
∑ (yi − y^i ) × x1i = 526 + 734.4 + 342.48 = 1602.88
​ ​ ​ ​ ​

i=1

∂J
Step 5: Compute the Partial Derivative ∂w ​

1 ​

The gradient of the cost function with respect to w1 is given by: ​

https://chatgpt.com/c/66f9eccd-6fe8-800e-95ae-f1014c8f2289 3/4
9/30/24, 5:48 AM ChatGPT
N
∂J 1
= − ∑ (yi − y^i )x1i
∂w1 N i=1
​ ​ ​ ​ ​ ​ ​

Plugging in the values:

∂J 1
= − × 1602.88 = −534.2933
∂w1 3
​ ​

Answer:

The partial derivative of the cost function with respect to w1 is approximately −534.29. ​

Final Summary:
[a] Using the trained linear regression model, the predicted time to failure for the machine
is 4.8 months.
∂J
[b] The calculated partial derivative ∂w is −534.29, indicating the direction and magnitude by
​

1 ​

which w1 should be adjusted to minimize the cost function

​

https://chatgpt.com/c/66f9eccd-6fe8-800e-95ae-f1014c8f2289 4/4