Partial Derivative - Chain Rule

what is the chain rule ?

consider the function y = m * x +b

i am going to work on the function which we can use for our machine learning problems .

consider a loss function h = [y - y_hat] ^2 which is also called mean square error

y - dependent on x

y_hat - predicted value

h = called loss

so why again two functions , ok

h = [m * x +b - y_hat]^2 looks complicated right ,this is where we will use the

chain rule ,how then ?

say partial derivative of m , b are

[1] dy /dm = x , dy/db = 1

do a substitution for h = u^2 , u = [y- y_hat]

[2]

dh/du = 2u = 2 * (y - y_hat)

du/dy = 1

so dh/dy = dh/du * du /dy --------[chain rule]

= 2 * (y - y_hat) *1

dh/dy = 2 * (y - y_hat)

so what is partial derivative of dh/dm and dh/db then ?

chain rule again from 1 and 2

dh/dm = dh/dy * dy/dm

[3] dh/dm = 2 * (y - y_hat) * x

dh/db = dh/dy * dy/db

[4] dh/db = 2 * (y - y_hat) * 1

values 3 and 4 are used in values corecction for m and b

new value of m = m - dh/dm

new value of b = b - dh /db

OK i think this is too much for now will do a implementation in our machine learning sample for better understanding .

Comments

SHA-256 initial values

The simple workout to arrive at the initial values for sha-256 The first 32 bit of the fractional part of the sqroot (first 8 prime number 2-19) Alright what does it say Sqrroot(prime)- Let’s say the first prime is 2 Sqroot(2) = 1.414213562373095 Convert to hexadecimal- Since we are worried about the fractional part alone Converting the fractional part would be easy Fractional part- 0.414213562373095 Multiply the fractional part with 16 to arrive at hex 0.414213562373095*16= 6.62741699796952 0.62741699796952*16= 10.03867196751232 0.03867196751232*16=0.61875148019712 0.61875148019712*16=9.90002368315392 0.90002368315392*16=14.40037893046272 0.40037893046272*16=6.40606288740352 0.40606288740352*16=6.49700619845632 0.49700619845632*16=7.95209917530112 Resulting hexadecimal would be 6a09e667 which is h0 := 0x6a09e667 Iam going to stop at the 8th iteration , why is that ? Since we are interested in 32 bit (8*4=32) Alright to make it clear Convert hexade...

Linear Regression with one variable - Introduction

It is not but making a some how clear relationship among variables the dependent and independent variables. talking in terms of maths the equation can be used meaningfully for something may be to determine /predict values from data. if y = m * x + b the values for m , b can be anything but has to appropriate to predict y so the loss which is difference from existing to prediction is close to zero ~0 to start with we can say the one variable as -x in some scenario m , b are called variables the equation stated about is a line equation we have any equation y = 2*x y = x*x y = 2x +2x*x so why the need of all these equations , it is all about playing data now a days in machine learning problems we create a data sets , lets consider as x y to be a value of x the datas . y = datas when we express the data as a function and plot in the graph we get the curves take some random data x and plot x and y x =1 , 2, ...

Running node in browser-https://stackblitz.com

Here comes the web browser development for web apps you can run your node project in web and do development as well please tryout the node or any other web development using https://stackblitz.com/ https://stackblitz.com/ run your scripts in the terminal Import your project from git hub and enjoy web over web development. see you soon.

Machine Learning , IOT , Microservices , python

Search This Blog