What Exactly is Machine Learning? A Definition - TheHackTech

What Exactly is Machine Learning?

What is Machine Learning?

"What is Machine Learning?" The short answer to the question is Netflix

When I describe the concept of machine learning, then I just ask whether Individual uses Netflix for entertainment. The answer is always yes. So how does Netflix "Just know" which movies you like or enjoy? This is called machine learning.

Talking technically, learning the machine is all about testing, testing, Test This training starts with data (your historical purchase behavior in Netflix) and then it applies the forecast of what you can like When you make an approximate selection, your answer is very much appreciated. And so is your fall.

Machine learning is a subfield of computer science that has evolved from the study of patterns In Artificial Intelligence and Computational Learning Principles by wisdom. machine learning
Study and investigation Learn from algorithms and Make predictions on the data. And all this happens automatically. 

Actually, You one day I am one and I am day one The commitment of resources like After developers and administrators early stages.

One example of Netflix is ​​different Big data is popular in the community That "CalTech University" machine Learning video library "speaks Toward Netflix competition!

If you do not fret or be embarrassed Do not know what machine learning is. By the end of this article, you will this Professional and scientific areas are low More than 10 years old Think fast on this
Track primer to walk in your next Machine-maker Learning.

Check Machine Learning Study and construction Algorithms that you can learn from And make predictions on the data.

Digging deeper

To consider machine Learning, there must be three elements

  • A pattern exists. There is a pattern in the rating films on Netflix.
  • We can't pin it down mathematically. so we are using machine learning to "learn" from data.
  • We have data Machine learning always starts with data.

Your goal is to learn or Insights from that data Here's something
The amount of automation in this process. Instead of trying for divine knowledge Manually, you're applying something
Algorithms to help answer the questions being presented

And that automation would be a computer!

Machine learning is not just a machine.

This Journey requires You need to Make much smart Human decision Part of someone like process; this Only machines.

Has two main categories of Machine Learning.

Supervised Learning: Make Forecast using data. it's also
Predictive modeling is known as. I recently took an online course
Where it was presented "Ham" or "spam." Landscape Was training a spam filter Review incoming email and Determine whether this was non-spam ("Ham") and should be kept in You know the rest of your inbox story. Another type of email Was spam and was kept in the Trash Folder. This goal is to make a prediction Specific Results One more scene Primary goal of To create supervised education
Models that are "normalized." this Predicts future accuracy Instead of the past.

Unsupervised Learning: This is the process of removing the structure From data and learning Best current data for example,
You can segment (more) to students Potential students) in groups
Which displays the same behavior. 

You Can get three groups Students performing Behavior (for each group) For example two years degree Students may be demanding The tenant's position as the predictor Variables; Masters can be students Is a homemade trend). But it is not variable Between two groups majority of Importantly, there is no right answer:

How does Machine Learning "Work"?

First of all, you train to learn a machine A model with training data (also known as As label data). This is the data in which Labeled with the result (is Email ham or spam?) It is called Model training because of the model Learning the relationship between Features of the data and its results

Such a relationship, using our email In the example, the length of text may be included Telegram, specific text in an email, Message length, etc...

Second, you make a prediction New data for which labels are
Unknown. It's actually easier than It seems. A new piece of email comes up.

"Machine" has not seen this piece First email. It predicts that this
The piece of email "hmm" is not spam Based on everything it knows.

Ground-level Model is learning Using the input from previous examples Output and then implement it In order to learn future input Predict future output. Is that correct?

This is not your father's data analysis!

For an indication of the complexity of the data, Modeling provides these questions Insights

  • How do I choose which features are Of my data for inclusion in Ideal?
  • Which model do I choose? Use?
  • How do I optimize this model? best performance?
  • How can I ensure that I am building Models that will be normal Undiscovered data?
  • Can I guess how good my model is 
  • There is a possibility of showing on ignore Data?

Of course, your mileage will vary As the answer to those questions, every The situation is unique. The mantra of Testing, testing, and testing applies here.

Writing about learning for one Viewer of professional teachers
At first, will be blush Nearly elementary if not aggressive.

But a quick review of the basics Always makes sense. Organ
Else in class, The machines are the same inside:

Input, output and target function Basically, the function statement
Creates output based on input.

In the case of education, the input is Student application Output is
A score that determines whether Students will enroll and will

The function is to score Student leadership. Now multiply by
Thousands of inputs (even billions) According to some data providers).

As a result of this activity, a hypothesis occurs This is the learning algorithm machine learning. Multiple algorithms A hypothesis would be set. 

Thought As it is a leadership scoring algorithm For students who are three different Degree categories: associate, Graduate, Master This is simple. 

But a layer is becoming deep, Bottom line is a threshold value. If the Student Lead Score is more than one Minimum, lead is approved and Followed by an entry Representative. This can be charted as A scatter picture (a common way Current machine learning).

This journey requires you to make many smart human decisions as part of the process; it’s not only machines.


Hi'i'm Rahim Ansari ,from India, I Love to Blogging, Desing Website, Web Developing and Desiging I Like to Learn and share Technical Hacking/Security tips with you,I Love my Friends.

Please Ask Question on Comment Box

Related Posts

Next Post »


27 August 2019 at 04:49 delete

any case, it is costly, sets aside a long effort to create, and needs adaptability because of the trouble in accomplishing consistency and accord for a typical information model for the whole association. artificial intelligence training in pune