i want to learn machine learning basics

Lesson-3

I want to learn machine learning basics

curious to know how to start learning machine learning I am excited too to tell you but first please get acquainted with these terms before we move on...

1. ML ➛ machine learning.

2. Labels ➛  in ml we use labels a lot for the thing we're going to predict for example:

the future price of a house, the kind of animal shown in the picture, etc.

in a simple linear regression, we denote it on the y-axis.

3. Features ➛ we all heard about the term feature it can be in terms of a smartphone feature or anything else. in ml we also have features but they somewhat differ for example in a spam detector the features could be:

➤words in the email text.

➤senders address.

➤time and date when an email sent.

so in ml, we give these features as input to our algorithm.

in a simple regression, we denote it on the x-axis.

4. Model ➛ a model defines the relationship between features and labels

for example in a spam detection model to label email spam or not spam we should have certain features associated the each of the labels then only we categorize them.

5. Dataset ➛ dataset or data is the most important part of our ml journey so in ml data is basically arranged in the form of tables or you may think these tables as an excel sheet. in ml we never use the full data instead we classify it into training and testing data with a 30-70% ratio.

6. Training data ➛ the data which we provide to our model to learn from is called training data.

for example, if we want our model to tell whether the given fruit is banana or apple. for that at first, we should have to teach/train it that  If the shape of an object is a long curving cylinder having color Yellow then it will be labeled as –Banana and if  If the shape of the object is rounded and depression at the top having color Red then it will be labeled as –Apple.

7. Testing data ➛ the data which we provide to our model to test if it is giving the desired result or not

for example, we show our model an apple and ask what is it?

the model can say it is apple but what if we show it a grape and ask what is it? if the answer is unknown fruit them fine but if it says it's an apple or banana then we need to train it with more features.

Bonus tip: always divide a dataset into two parts 70% for training and rest 30% for testing.

8. Instance ➛ instances are nothing but the rows in the dataset.

9. Prediction ➛ after training when we test our model with an unknown fruit then whatever be the output it is called prediction.

10. Overfitting ➛ supposes we oversight the  70% of dataset for training concept and we used a full 100%. to train our model.

now we didn't left with any training dataset and if we want to train we have the use the data which is well known to the model when we try to test the data on training dataset it always gives 100% accuracy but in-depth we also know our results are biased and results may differ if an unknown test data is provided to the model.

for example- before the exam, if the teacher tells the students what questions are coming in exams so what's the purpose of even conducting a test.

Any doubts or questions then comment below ⇓⇓⇓ 

Thanks!!! for Visiting Asaanhai or lucky5522  😀😀