0% found this document useful (0 votes)
4 views28 pages

12-How To Use Python For Machine Learning

Uploaded by

Kaan Karapınar
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
4 views28 pages

12-How To Use Python For Machine Learning

Uploaded by

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

< 1 />

YZV102E
How to Use Python for Machine Learning?

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 2 />

Week 11
How to Use Python for
Scientific Computation?

Machine Learning Concepts

Regression

Classification

Unsupervised Learning
Machine Learning Concepts
Reinforcement Learning

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 3 />

What is Machine
ARTIFICIAL INTELLIGENCE
Learning?
• Application of algorithms that learn from
examples (data)

• Representation and generalization


MACHINE LEARNING
• Useful in everyday life

• Especially useful in data analysis

DEEP LEARNING

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 4 />

Rules
Classical
Answers
Programming
Data

Answers
Machine
Rules
Learning
Data

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 5 />

Example Applications
• Computer Vision • Object Classification in
• Robotics Photos
• Virtual Assistants • Image Caption Generation
• Email Spam Filtering • Automatic Game Playing
• Product Recommendation
• Self-driving Cars
• Machine Translation

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


Machine Learning Styles
< 6 />

Unsupervised Supervised Reinforcement


Learning Learning Learning

The pattern must be found. The pattern is given. Policy optimization.

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


7
< />

How to Build Machine Learning


Systems
Labels
Training New
Dataset Learning
Final Model Data
Labels Algorithm
Validation Dataset

Raw Test Dataset


Data Labels

Preprocessing Learning Evaluation Prediction

• Feature Extraction & Scaling • Model Selection


• Feature Selection • Cross-Validation
• Dimensionality Reduction • Performance
• Sampling Metrics
• Hyperparameter
Optimization
Algorithms
< 8 />

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


Sample Dataset
< 9 />

Attributes
1. age
2. sex
3. chest pain type (4 values)
4. resting blood pressure
5. serum cholestoral in mg/dl
6. fasting blood sugar > 120 mg/dl
7. resting electrocardiographic results (values 0,1,2)
8. maximum heart rate achieved
9. exercise induced angina
10. oldpeak = ST depression induced by exercise relative to rest
11. the slope of the peak exercise ST segment
12. number of major vessels (0-3) colored by flourosopy
13. thal: 3 = normal; 6 = fixed defect; 7 = reversable defect

[Link]

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 10 />

•Simple and efficient tools for predictive data


analysis
•Accessible to everybody, and reusable in various
contexts
•Built on NumPy, SciPy, and matplotlib
•Open source, commercially usable
• BSD license

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 11 />

Week 11
How to Use Python for
Scientific Computation?

Machine Learning Concepts

Regression

Classification

Unsupervised Learning
Regression
Reinforcement Learning

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


Linear Regression
< 12 />

1 1 !
𝑀𝐴𝐸 = ' 𝑦 − 𝑦* 𝑀𝑆𝐸 = ' 𝑦 − 𝑦*
𝑛 𝑛

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


Linear Regression Sample Dataset Diabetes Data 13 < />

[Link]

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


Linear Regression Sample Dataset Diabetes Data 14 < />

import numpy as np
from sklearn import datasets, linear_model
from [Link] import mean_squared_error, r2_score

# Load the diabetes dataset


diabetes_X, diabetes_y = datasets.load_diabetes(return_X_y=True)

# Use only one feature


diabetes_X = diabetes_X[:, [Link], 2]

# Split the data into training/testing sets


diabetes_X_train = diabetes_X[:-20]
diabetes_X_test = diabetes_X[-20:]

# Split the targets into training/testing sets


diabetes_y_train = diabetes_y[:-20]
diabetes_y_test = diabetes_y[-20:]

# Create linear regression object


regr = linear_model.LinearRegression()

# Train the model using the training sets


[Link](diabetes_X_train, diabetes_y_train)

# Make predictions using the testing set


diabetes_y_pred = [Link](diabetes_X_test)

[Link]

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


Feature Engineering
< 15 />

• Dimension Reduction
• Derived variables
• Date -> Year, Month, Day
• Create one or more variables
based on other variable(s).
• Dummy Variables
• One-hot-encoding
• Label Encoding
• Turning Integers to Categories

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 16 />

Week 11
How to Use Python for
Scientific Computation?

Machine Learning Concepts

Regression

Classification

Unsupervised Learning
Classification
Reinforcement Learning

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 17 />

Classification – The Famous IRIS Dataset

•150 samples
•3 labels: species of Iris (Iris setosa, Iris
virginica and Iris versicolor)
•4 features: Sepal length,Sepal width,Petal
length,Petal Width in cm

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


18
< />
SPAM Mail Classification
IRIS Dataset
Credit Card Fraud Detection
Face Recognition

Binary Classification Muliclass Classification

Number of target labels

Classification

Number of labels for each example

Single Label Multilabel

SPAM Mail Classification News category identification


Face Recognition

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 19 />

K-NN Classification

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 20 />

from sklearn import datasets


iris = datasets.load_iris()

x = [Link]
y = [Link]

from sklearn.model_selection import train_test_split


x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=.5)

from sklearn import neighbors


classifier = [Link](n_neigbors=5)

[Link](x_train, y_train)

predictions = [Link](x_test)

from [Link] import accuracy_score


print(accuracy_score(y_test, predictions))

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 21 />

Decision Tree Classifier

From Alpaydin, 2010

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 22 />

from sklearn import datasets


iris = datasets.load_iris()

x = [Link]
y = [Link]

from sklearn.model_selection import train_test_split


x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=.5)

from sklearn import tree


classifier = [Link]()

[Link](x_train, y_train)

predictions = [Link](x_test)

from [Link] import accuracy_score


print(accuracy_score(y_test, predictions))

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 23 />

Week 11
How to Use Python for
Scientific Computation?

Machine Learning Concepts

Regression

Classification

Unsupervised Learning
Unsupervised Learning
Reinforcement Learning

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


Unsupervised Learning
< 24 />

Explore structure of data to extract meaningful information without the


guidance of a known outcome variable or reward function.
x3
• Clustering organizes data into meaningful subgroups • Dimensionality Reduction
(clusters) – “unsupervised classification” compresses data onto smaller
dimensional subspace while
retaining most of the relevant
x1 information
x2
x1
z1

z2

x2
YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ
< 25 />

Week 11
How to Use Python for
Scientific Computation?

Machine Learning Concepts

Regression

Classification

Unsupervised Learning
Reinforcement Learning
Reinforcement Learning

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


< 26 />

Classification – The Famous IRIS Dataset


Building an agent that improves its performance based on
interactions with the environment by optimizing a policy.

An agent uses reinforcement state

learning, through interaction Agent


with the environment, to learn
reward
a series of actions (inputs) that action
maximizes the reward signal, Environment
measured by a reward function.

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ


27
< />
< 28 />

End of chapter
Please study yourself on what you have learned today!

YZV102E Introduction to Programming for Data Science (Python) by A. Cüneyd TANTUĞ

You might also like