# Machine Learning Algorithms — Linear Regression in Hindi

--

Machine Learning Algorithms — **Linear Regression in Hindi** — Is post me hum bat Karne wale hai linear regression k bare me. Jab bhi hum machine learning ke algorithms ko padhna start karte hai to hum start linear regression algorithm ke saath karte hai. So, aaj iss post me hum isi algorithm ke baare me baat karenge.

To linear regression kya hota hai? Simple ya multiple linear regression ka equation hota hai y = mx + c. Isi equation ka use karke hum best fit line ko find karte hai. Chaliye ab linear regression ke definition ke bare me jante hai.

Read More — What is Generator and Yield in Python

# Linear Regression in Hindi

Linear regression best fit line ka use Karke independent aur dependent variable Ke beech me relation establish Karta hai. Yaha par dependent variable Y hai aur independent variable X.

**Y = mx + c** linear regression k equation ko dikhata hai. Yaha pr **m** slope ko represent karta hai aur **c** ek constant value ko, jisko hum line ka intercept bhi bolte hai.

Chaliye ab iss algorithm ko hum ek example ke help se samajhte hai. Let us suppose ki hum kisi foreign university me admission lene ja rahe h. Hume admission lene ke liye ek achche GRE score ki jaroorat Hoti hai.

Isi ke Saath old universities ki rating, extracurricular activities and grades ki Bhi jaroorat hoti hai. To yaha par humara goal hai admission ke chances ko find Karna. Yaha par chance of admission dependent variable hai and university rating, gre score, extracurricular activities independent variable hai. Yaha par humara admission in sub factors par depend Karta hai isiliye ye dependent variable hai.

Assumption ka matlab hota hai imagination yaani ki kalpana karna.

- Kisi X value ka liye y normally distributed hota hai. Dusre word me ye keh sakte hai ki data normally distributed hota h.
- Variable x aur y linearly dependent hote hai.

# Overfitting and Underfitting In Hindi

Overfitting ka matlab hota ki humara model training data par achcha perform karta hai magar test data par iska performance achcha nahi hota.

Underfitting ka matlab hota hai ki, humara model training aur testing dono par hi achcha perform nahi karta.

So, humara question ye hai ki, Kya linear regression model overfit our underfit hote hai. To iska answer hai yes. Linear regression ke overfit hone ke chance hote hai.

# Advantages of Linear Regression — Hindi

- Jo data linearly separable Hote Hai Unke Saath ye achcha perform krta Hai.
- Hum linear regression model ko bahut easily train kr Skte hai.

# Disadvantages

- Inki performance outliers ke vajah se affect hoti hai.
- Isme feature scaling ki jyada jaroorat padti.
- Linear regression model me overfitting jyada hoti Hai.
- Agar dependent aur independent variable me correlation hote hai to model achcha perform nahi krta Hai.

# How to measure the performance of Linear Regression model in Hindi

Linear regression ke performance ko hum bahut se tariko se measure kr skte Hai.

# Mean Square error

Mean Square Error Hume ye batata hai ki jo humara best fit line hai wo data points ke kitne close hai. ye predicted points aur actual points ka difference nikalta hai usko jodta hai aur fir unka avergae nikalta hai.

# Root Mean Square Error

RMSE ka formula MSE ke jaisa he hota bus issme antar ye hota hai ki ye square root nikalta hai.

R-squaed ye batata hai ki data regression line ke kitna close hai. Isse hum coefficient of determinent bhi kehte hai.

R-square ki value Humesa 0 se 100% ke beech me Hoti Hai. jitna jyada iski value hogi utna he Achcha ye model performs Karega.

*Originally published at **https://code4hub.com** on January 12, 2021.*