Machine Learning

Machine learning can be used to solve a wide variety of problems by discovering patterns in large datasets and making predictions on new data. Some examples of machine learning applications include recommendation engines used by Netflix, Spotify, Amazon to make personalized suggestions, algorithms used by Uber to estimate transportation times and costs, and technologies used in apps like Google Photos, Siri, and Google Camera. Python is often considered the best programming language for machine learning due to its excellent scientific libraries like NumPy, Pandas and scikit-learn, as well as its compatibility with popular deep learning frameworks like TensorFlow.

Uploaded by

Kajal

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

34 views

Machine Learning

Uploaded by

Kajal

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 5

UNIT 4

What kind of problems can be solved using Machine Learning?

Machine learning is an important area of artificial intelligence.

The objective of machine learning is to discover knowledge and make intelligent decisions.
Machine Learning is an application of Artificial Intelligence and
is revolutionizing the way companies do business. At it’s core, it’s
an algorithm or model that learns patterns in big data and then
predicts similar patterns in new data.
In layman’s terms, it’s the theory that machines should be able
to learn and adapt through experience to produce reliable,
repeatable decisions and results.

You probably use it dozens of time a day

without even knowing it.
There are many examples of machine learning. Just to give a few examples:
1. Netflix - Recommendation engine (Creating movie suggestions based on
viewing history)

2. Spotify - Recommendation engine (Songs suggestion)

3. Uber - Machine Learning Algorithms (For predicting time taken to reach
destination, time of arrival, estimated cost, etc)
4. Amazon - Recommendation engine (Shopping suggestions)
5. Facebook/Twitter/Social media sites in general - Curated timeline

6. Google Photos app: It can recognize scenery, people etc in images.

7. Siri: Uses Neural network, speech recognition and machine learning models.
8. Google's new camera app: This one can accurately take photos in dark and
colors them as if they were taken in light.
In Big data, the gathering of datasets is so huge Machine Learning algorithm is incorporated for
and composite so out-dated learning the processing of high volume of data.
techniques are not suitable for them

Some modern learning techniques are greatly desirable for resolving the big data problem.
Which programming language should I use for machine learning?

In my subjective opinion, the best programming language for machine learning in 2017
is Python.

Here are some of the things Python offers that makes it such a great language for machine
learning:

• Excellent scientific libraries. SciPy, pandas, NumPy and many other

libraries make Python a prime ecosystem for any scientific computing.
• scikit-learn. In my opinion, the best machine learning library out there.
• Unparalleled compatibility with popular deep learning
frameworks. TensorFlow, Theano, Keras, CNTK and MXNet all come
with Python APIs.
• Reasonable speed when programmed efficiently. Python is certainly
much faster than R without any additional scaling. NumPy allows for
efficient linear algebra without any hassle. This is very important in
machine learning.
• The best syntax. This is obviously subjective, but most people will tell you
that Python arguably has the most pleasant syntax out of all programming
languages.
• A general purpose programming language. It can easily be integrated
with backends written in Python or even other languages. It’s a production-
level language, unlike R. This is its strongest advantage compared to the
latter.
• Excellent community support. There is a huge and thriving open source
community for machine learning applications in Python.
What is the most useful library for machine learning in Python?

Basic “must-know”:

• numpy (for fast operations on numbers)

• pandas (for data analysis and transformation)
• matplotlib or seaborn (for visualization)
• scikit-learn (for addressing common problems in data science like
clustering, metrics, dataset manipulation, etc.)
Then for more powerful stuff:

• tensorflow and/or pytorch (for more sophisticated operations on data -

forming and solving computation graphs)
• keras (nor add-on to tensorflow, for simpler interface to neural networks)
• scipy (for advanced mathematical and numerical operations)