Regarding how the equation -64 + 2 * hours came: In case of the single variable equation like y = a+ b * x, if we decide to use it say (This assumption is important because We can also take another equations of single independent variable x which contain its polynomial terms like a+b*x + c*x^2 for more complex decision boundaries [for classification] which are not straight lines) Usually the user of the algorithm will start with some random constants like a = 0, b = 0. Based on the given training data which would contain the values of x, y for multiple training examples, An optimization algorithm (like gradient descent, quasi newton algorithms etc) would be used to minimize the cost function (which is a function of the training data and the constants [which are formally called parameters] a, b). [Training data would contain multiple (x,y) pairs in our case where x is the input and y is the expected output for that input] Minimizing the cost function with respect to a, b would give those values of a, b for which cost function got minimized, in other words those a, b which fit the training data well enough and give good results on the training data. In such a way a, b are found or if I want to sound "fancy", this is the way a and b are "learnt" or if I want to sound "extra fancy", this is usually the way a and b are "fitted" (whole process called training or parameter fitting) This is how the equation is found in actual applications/programs which use logistic regression. In exams, probably the equation would be given for such questions.
your way of teaching is simple awesome sir,but i have on a doubt in the above problem i.e, how to calculate the z value and threshold value kindly clarify it sir ,thank you.
![Logistic Regression [Simply explained]](/img/1.gif)
Regarding how the equation -64 + 2 * hours came:
In case of the single variable equation like y = a+ b * x, if we decide to use it say (This assumption is important because We can also take another equations of single independent variable x which contain its polynomial terms like a+b*x + c*x^2 for more complex decision boundaries [for classification] which are not straight lines)
Usually the user of the algorithm will start with some random constants like a = 0, b = 0.
Based on the given training data which would contain the values of x, y for multiple training examples, An optimization algorithm (like gradient descent, quasi newton algorithms etc) would be used to minimize the cost function (which is a function of the training data and the constants [which are formally called parameters] a, b).
[Training data would contain multiple (x,y) pairs in our case where x is the input and y is the expected output for that input]
Minimizing the cost function with respect to a, b would give those values of a, b for which cost function got minimized, in other words those a, b which fit the training data well enough and give good results on the training data.
In such a way a, b are found or
if I want to sound "fancy", this is the way a and b are "learnt" or
if I want to sound "extra fancy", this is usually the way a and b are "fitted" (whole process called training or parameter fitting)
This is how the equation is found in actual applications/programs which use logistic regression. In exams, probably the equation would be given for such questions.
your way of teaching is simple awesome sir,but i have on a doubt in the above problem i.e, how to calculate the z value and threshold value kindly clarify it sir ,thank you.
There were, not any particular value for z is calculated,
One equation will be given to you in the question
Your videos are very informative and well designed. One question on this, "model is suggested by optimizer..." which optimizer you're talking about?
I have doubt sir..how to calculate equation if it is not given
From where you ot equation y = -64 + 2*hours or can we build this from the data set given that is my question
Did u got the answer?
Yes specify it I did not get it
@@yasmenabdullah8283 yes, the eqn will be given in ques
@@GazalKhan-x8e will be given
@@yasmenabdullah8283 no
How the z value is calculated -64+2*hours
Please clarify my doubt
It's a function it will be given in the question.
@@rhealivesonabeach9145 if it's not giving in the question then how to find it out
Awesome content! Keep it up. Thank you sir.
Welcome
Do like share and subscribe
One of the best teachers
Thank You
Do like share and subscribe