Thursday, October 2, 2014

Data Mining Topics for Digital Marketing

• Mining for Ad Relevance and Ranking
o Ad relevance measurement
o Ad ranking algorithms
o Ad text creation and evaluation
o CTR and conversion rate prediction
o Real-time bidding optimization
• Audience Intelligence & User Modeling
o User Tracking
o Understanding user intent
o Modeling online user behaviors for targeted advertisement
o User segmentation and profiling
o Demographics & location prediction
o Personalized advertising
• Content Understanding and Content Marketing
o Content-targeted advertising
o Opinion/sentiment mining
o Web scale information extraction for online advertisement
o Text mining techniques such as named entity extraction, query classification, keyword extraction, and other topics
o Understanding multimedia content for online advertisement
• Social, Mobile Advertising, and new advertising channels and formats
o Advertising through social networks and microblogging (such as Facebook and Twitter)
o Advertising through deals (such as Groupon and LivingSocial)
o Advertising on new channels such as mobile devices
o Mobile advertising
o Video advertising
o Native advertising
o Viral marketing
• Advertising Ecosystems
o Auction theory in online advertising
o Demand and supply volume prediction
o Measurement of online advertising effectiveness
o Search Engine Marketing, Optimization (SEMs, SEOs)
o Systems and technologies in ad exchange and RTB
• Trust and Privacy
o Consumer privacy and data use policy
o Privacy preserving data mining approaches
o Fraud and spam detection & prevention in online advertisements

Sunday, September 21, 2014

The future of AI and deep learning

Gigaom is hosting a meetup tonight in San Francisco about artificial intelligence and deep learning, and while the event itself is sold out I’m happy to announce that we will be streaming the event live. You can watch it, starting at 6 p.m. Pacific Time, here: http://new.livestream.com/gigaom/FutureofAI

Saturday, August 16, 2014

Big Data - from Simple Ideas to Advanced Concepts

Basic ideas:

Big Data - explained in a fun and easy way



5 V of Big Data
Basic Principles of Big Data System


More information at refer links:

Advanced concepts:

We, the human see data (structured), create new data (unstructured and structured) and the demand is finding the relationship inside new data. That's why big data was born !


How ?


Tuesday, July 29, 2014

Tại sao công nghệ AI deep learning không chỉ dành cho Facebook, Google, IBM, Netflix


Mục đích: Hiện thực 1 Java Web Service, nhằm đưa ra những thông tin hữu ích từ logs, có giá trị dựa trên những gì bạn thích hoặc tiềm năng trong tương lai, hoặc gây ra rủi ro nguy hiểm cao để mỗi cá nhân tự phòng tránh. (dành cho nhu cầu cá nhân mỗi người)




Implemented code:
Dùng Java 8 with Lambda
https://bitbucket.org/trieunt/rfx/src

1 vài open source tham khảo:
http://deeplearning4j.org/
http://jfuzzylogic.sourceforge.net/html/index.html
https://github.com/orientechnologies/orientdb/wiki/Graph-Database-Tinkerpop

Useful links:
Make suggestions based on what you actually like about your favorite information experience
http://gigaom.com/2014/07/29/robots-helped-inspire-deep-learning-and-might-become-its-killer-app/
http://blog.echen.me/2011/07/18/introduction-to-restricted-boltzmann-machines/
The application of Deep Learning in Collaborative Filtering
Netflix Is 'Training' Its Recommendation System By Using Amazon's Cloud To Mimic The Human Brain
http://www.quora.com/Deep-Learning/Whats-the-most-effective-way-to-get-started-with-Deep-Learning
http://ufldl.stanford.edu/wiki/index.php/UFLDL_Tutorial
http://deeplearning.stanford.edu/wiki/index.php/Main_Page
http://www.socher.org/index.php/DeepLearningTutorial/DeepLearningTutorial

Practical case studies
http://danielnouri.org/notes/2014/01/10/using-deep-learning-to-listen-for-whales/
http://radimrehurek.com/2013/09/deep-learning-with-word2vec-and-gensim/
http://radar.oreilly.com/2014/07/how-to-build-and-run-your-first-deep-learning-network.html

Tuesday, July 22, 2014

Emerging Trends in Big Data Technologies

Emerging Trends in Big Data Technologies
  • Storm: Apache Storm is an open source distributed real-time computation system. Storm makes it easy to process streams of data, doing for real-time processing what Hadoop did for batch processing.
  • Spark: Spark is an in-memory data-processing platform that is compatible with Hadoop data sources but runs much faster than Hadoop MapReduce. It’s well suited for machine learning jobs, as well as interactive data queries, and is easier for many developers because it includes APIs in Scala, Python and Java.
  • Apache Hive: Apache Hive facilitates querying and managing large datasets residing in distributed storage. It also allows the map reduce programmers to plug in custom mappers and reducers.
  • Apache Tajo: Apache Tajo is a big data relational and distributed data warehouse system for Apache Hadoop. Tajo is designed for low-latency and scalable ad-hoc queries, online aggregation, and ETL (extract-transform-load process) on large-data sets stored on HDFS (Hadoop Distributed File System) and other data sources.
  • Twitter'Summingbird
Full report: http://www.infoq.com/research/big-data-emerging-trends

Friday, July 11, 2014

Khoa học dữ liệu, triết học và bóng đá (How Big Data Helped Germany in the World Cup 2014)


Tóm tắt bằng Vietnamese như sau:
Sau thất bại (2 lần hạng 3 thế giới, 2 lần vào chung kết Euro) từ 2006 (cuộc cách mạng bóng đá Đức do Jürgen Klinsmann đề xuất), đội tuyển Đức đã kết hợp giữa khoa học và thể thao một cách chặt chẽ nhằm tìm kiếm  danh hiệu thứ 4 (vô địch World Cup 2014).
Theo như trợ lý huấn luyện Hansi Flick, trong 2 năm qua, các sinh viên ở truờng đại học thể thao Cologne đã phát triển một hệ thống gồm cơ sở dữ liệu về cầu thủ, cách sử dụng chiến thuật và cách đá, đi bóng của từng câu thủ.



Vì vậy, dữ liệu của 736 cầu thủ đá ỏ World Cup (gồm luôn tuyển Đức) đã đuợc số hóa để giúp ban huấn luyện đề ra từng chiến thuật cụ thể cho từng trận đấu ở giải kỳ này.
=> trận hòa Ghana chắc do thiếu data #_#

Câu hỏi: Với mỗi trận đấu, sẽ tồn tại ít nhất một cách để phá lối đá của đối phuơng và phát huy tối đa sức tấn công để ghi bàn ?
chờ đợi trân chung kết để xem điều này đúng hay sai :)


SAP data at HoffenheimThe data can be analysed in real-time by data experts - and training schedules can be adapted
NSA , a super analytic in football , is developed at Germany for this World Cup 2014
“The sports students in Cologne have been studying in great detail our opponent and put every play they’ve run, every newspaper article on them, and everything about them out there under the microscope and made all that data available to us,” 

Read more: http://www.dailystar.com.lb/Sports/Football/2014/Jul-08/263019-university-boffins-steer-german-tactics.ashx#ixzz379AoFIWj
(The Daily Star :: Lebanon News :: http://www.dailystar.com.lb) 


big data is about capturing the "moment". The German team was able to capture and analyze each and every moment about the Brazilian team including the passes they play, how they react upon pressure, and even every quote about them in newspapers. They were able to analyze hidden team behaviors and strategy information.
From https://www.linkedin.com/today/post/article/20140709143632-54591340-how-big-data-helped-germany-break-brazil-s-hearts-in-the-world-cup
Big Data is about capturing every single “moment” throughout the human life
https://www.linkedin.com/today/post/article/20140627113503-54591340-big-data-utopia

http://blogs.bridgei2i.com/2014/06/25/for-the-love-of-soccer-and-data-analytics/