Scribd
Upload
12 views2 pages

Machine Learning

Uploaded by

Satyam Lamsal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
12 views2 pages

Machine Learning

Uploaded by

Satyam Lamsal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 2

Notes:

1. Supervised learning

Problem 1:
i.e price vs size plot study
create a function from given data
and that helps to predict future input
(Regression Problem)

Problem 2:
plot of
i.e
size of tumer vs malignantOrNot (1,0)
From the above data set we predict future input for the size of tumer
(Classification Problem : if the output is discrete)

Problem 3:
i.e
tumer size vs age of person and corresponding circles(+) and cross(-) for disease

breast cancer problem(use a lot of features)


thickness, uniformity of cell shape,size , adhesion ....

learn an algorithm that uses a lot of input


i.e
algorithm called support vector machine that supports inf input about an object
kernels

(All Fall Under Supervised Learning)


During training both X and Y are giving an algorithm is created that
maps the value of X with Y creating a function that predicts future X for
respective Y

Machine Learning Strategies ::


--> Make ML projects systematic and productive

Deep Learning:(hot right now)


-->

after the above things::

Unsupervised learning::
-->
1. Just given labels i mean values like X and find interesting
facts and infos of that data
2. I.e grouping similar datas like grouping News related to tech in one
place
3. Genetic Data analysis,
4. Clustering algorithms (i.e creating community in social network)
5. grouping galaxies
6. Unlabeled data and finding intresting things about it

7. Cocktail Party Problem


i.e Separte mic sound of different people for noise
cancelling ||| ICA , Independent Components Analysis
8. Learn analogies from unlabeled data

Reinforcement Learning :
--> Think of it as training a dog
--> Let it do whatever wants and judge good or bad .. after a long time it becomes
much good than bad.
--> let a robot do any work and comment on its performance good or bad

You might also like