Course Web Site. I suggest you to watch online course Analytics in Python

0. Introduction

• Python and Libraries
• Machine Learning
• Data Science: Getting Data From Twitter

1. Compare Basic Machine Learning Algorithms

• LinearRegression
• LogisticRegression
• Decision Tree
• Application: Introduction and test and compare different algorithms on synthethic data

2. Anomaly Detection Kaggle Kernel

• 3 Sigma Rule and link
• Isolation Forest
• Application : Google Trend Data

3. Intro 2 Text Mining

• TF-IDF explained
• Creating Text Mining Model
• Test on Twitter with pre-trained model

4. Text Summarization

• Algorithm to Extract Summary
• Application : Summary of Wikipedia page

5. Web Scraping

• Application : Berkay Öztürk's talk on UKitapScraper

6. Fundamentals of Machine Learning From Data

• Learning From Data
• Test on synthethic data

7. Midterm Questions
8. Predict political party based on votes

• Political Analysis
• Political Data Science

9. Supervised Learning: ANN

• Keras Deep Learning Library
• Recognize handwritten numbers

10. Advanced NLP

• wor2vec on Hurriyet
• Turkish NLP Library Zemberek
• Simple chatbot

11. Unsupervised Learning

• Kmeans and hierarchical clustering

12. Network Analysis

• networkx library
• Social Network Analysis with Python
• Centrality measures
• Game of Thrones

As course advances we will add more alternative projects. You must do at least one project. You can also propose a new project.

Below you can find the link for determining your project groups.

• http://bit.ly/CMPE251Projects18

Use this link to write the name of the project, your data source and your team mates and the name of your team.

The most critical part of your project is the correctness of your traninig labeled data. If your data is not good, you will receive very low points.

Data Collection You will get data from eksisozluk with web scraper. Each student will label 500 comments on eksi. 5 label means 5 class.

• 5 Very Positive
• 4 Positive
• 3 Neutral
• 2 Negative
• 1 Very Negative

Each group MUST have different data sources. Different "gundem" topics from Eksisozluk.

Machine Learning Use ML algorithms for Sentiment Analysis On EksiSozluk. Indicate your results.

Data Collection You will get data from Zaytung and normal newspapers websites with web scraper.

• Zaytung news: 1
• Normal newspapers: 0

Before you might need to look for irony detection

• irony detection in Turkish 1
• irony detection in Turkish 2

Machine Learning Use ML algorithms for Sentiment Analysis On EksiSozluk. Indicate your results.

Data Collection Use NRC Emotion Lexicon and Kaggle song lyrics dataset

Machine Learning Use textrank algorithm to create a new song lyric from a popular singer. Then use it to create a combined lyric of various singers.

we are defining a different relation, which determines a connection between two sentences if there is a “similarity” relation between them, where “similarity” is measured as a function of their content overlap.

Task Create new songs of some artists based on 2 algorıthms

• Simple text summarization
• Text summarization based on textrank algorithm

And compare the results.

2 datasets

• https://dataverse.harvard.edu/file.xhtml?persistentId=doi:10.7910/DVN/OPQMVF/GIPF3O&version=1.0
• https://dataverse.harvard.edu/file.xhtml?persistentId=doi:10.7910/DVN/OPQMVF/MTUJ5F&version=1.0

3 algorithms

• Isolation Forest
• Self-organizing maps
• Local Outlier Factor

Dear students,

I updated the time slots for CMPE 251 projects. Go to the following link below, to see the updated time slots

• http://bit.ly/CMPE251Projects18

You can also see here, updated time slots.

• Please come 10 minutes earlier than your presentation time!!
• Bring printed version of your project report with you
• Your project code&data&report&slides should be given within a CD-ROM & USB
• I will take the CD-ROM but not the USB