San Francisco Hacker News - Machine Learning for Hackers
2 likes1,310 views
This document discusses machine learning and data science. It explains that machine learning can be used to analyze large datasets with many variables, unlike traditional statistics which is better for smaller datasets with fewer variables. The document then provides examples of different types of data and how machine learning can be applied to images, text, time series data and more. It also defines the roles of data scientists and data engineers. Finally, it provides recommendations for learning machine learning including starting with basic analytics, understanding key concepts and using tools like scikit-learn, Spark and Hadoop.
San Francisco Hacker News - Machine Learning for Hackers
- 1. Machine Learning for Hackers is how we make sense of big data. Adam Gibson 2-27-2014 SFHN
- 2. BIG DATA & STATISTICS • Statistics – Group by, aggregate, count,average, mean,p values,mode,correlations, exploring, < 100 variables • Machine Learning – Label this image, Predict the next event, Pick out the anomalies – aka learn from data not count it, group data by similarities, > 100 variables.
- 4. Many kinds of data Wow. Unstructured Text SQL Video XML Images JSON Time Series CSV Data Scientists Structured We know this, and just process it.
- 5. WHAT do machines learn? • Machine learning is a general tool that can work with • • • • • • • various data types. Images = Machine vision Text = Natural-language processing Time-series = Prediction Facial recognition => Security Text => Customer profiles/Recommendation engines Time-series => stock-market trading platforms NLP => Customer service
- 6. WHAT IS A DATA SCIENTIST? Analyst Exploratory analysis of data, typically on smaller data sets. Understands the algorithms and interprets data. Distributed Systems Engineer Implements production data crunching, also known as the nosql person. They handle distributed systems and workloads, APIs, perhaps even data collection and storage
- 7. What kinds of Machine Learning Are there? Unsupervised – Clustering (group things that are similar, regression (correlation != causation ring a bell?) Supervised – Label all the things! Predict the future!
- 8. How does this affect me?
- 9. I will leave who does this to your imagination Ad Targeting Recommends you Movies Brings you search results Recognizes your face in the camera Drives your car Automatically disables your credit card when you leave the country
- 10. Can I do this? The shortcut here is to start with basics – for example google analytics, understanding churn rate. Pick up a more advanced understanding after that if it still seems interesting. If you are in to backends start with distributed systems, get your math basics up enough to understand what the guy on the other side of the table who's asking you to put the algorithm in to production is saying
- 11. Resources Coursera Machine Learning Reddit Machine Learning DataTau (hacker news for data scientists) More mathy Stanford Machine Learning
- 12. Tools
- 13. Analysts Data Engineers Spark http://scikit-learn.org/stable/ Julia Lang R Lang Hadoop Storm Hadoop QuickStart VM