This is the second lecture for popularizing Java machine learning. We introduce Weka, which provides UI and API so that both design and coding can be implemented. We have included cases that are completely suitable for practical application in the lecture.
How to apply Java machine learning in practice using weka
Adoption of optimal algorithm through comparative analysis
Decision-making basis derivation using only feature selection
Structured text mining such as Hangul survey
Proof of causal relationship through classification analysis after association analysis
Application of artificial neural networks for image analysis
Integration of weka and R programs
Want to implement machine learning in Java? 👉 Try using machine learning in your work with the intermediate level of Java Machine Learning Weka .
1. why (purpose)
The goal is to build a collaborative system that enables rapid decision-making using data. : Introducing weka, which enables both design and programming of Java machine learning.
2. what (lecture content)
This is a practical machine learning application case implemented solely with Weka. : We've adapted various application cases into familiar content. So, let's briefly introduce the content.
2.1 Adoption of the optimal algorithm by the experimenter : Selecting the optimal model through statistical testing of significance level (p-value) (Have you heard this before?)
2.2 Presenting the basis for decision-making based on feature selection alone : You can create decision-making information simply by selecting specific attributes. Integration with R programs is a bonus.
2.3 Korean Survey Text Mining : No more struggling with difficult Korean morphemes! Simple Korean surveys are possible with just the basic features.
2.4 Correlation and classification analysis of the 1984 U.S. House of Representatives election results : The Obama camp did not predict the election pledges, but rather selected the website to raise campaign funds through statistical analysis. : What's really important is knowing which promises are directly linked to getting elected, right?
2.5 Image Analysis Using Artificial Neural Networks and Image Filters :I've been waiting for a long time for dl4j to be in beta. : Introducing the built-in artificial neural network provided by Weka and wekadeeplearning4j.
2.6 Estimating course completion time through regression analysis : I used it to figure out how long I should delay the release of the lecture.
3. Method
The above process is explained in the following three steps.
3.1 Theoretical Explanation : I'll give you a brief background explanation. Really simply, just the essentials.
3.2 KnowledgeFlow Design : Weka's biggest advantage is that you can do machine learning without knowing programming.
3.3 Java Programming : Another advantage of Weka is that Weka provides everything for design and coding.
4. IF (GET)
You can apply data analysis loaded on traditional IT systems built on the Java platform. : A journey of a thousand miles begins with a single step. If you know how to analyze traditional IT, wouldn't you be able to analyze ICBMs well?
You will learn how to understand the real world with data. : The goal is to understand the previously invisible reality through data.
This is an intermediate course, improved from the beginner level, to popularize Java machine learning. : I have improved the beginner course through feedback from students and self-reflection.
6. Lecture environment
I use weka 3.9.3 on Windows OS. (3.9.4 has a bug when loading ANSI type files.)
7. Lecture materials
After registering for the course, in the second lesson of Section 1 (installing Weka software and downloading lecture materials) Click the cloud icon to download (55MB)
8. Survey
We are conducting a five-item survey to help establish the direction of our advanced Java machine learning course. We ask for your participation. http://bit.ly/javamachinelearning_survey
9. Continuation of the extra lecture
We will continue to upload ideas gleaned from students' valuable questions and useful information from other media as supplementary lectures.
I look forward to strengthening communication with my students and improving their satisfaction with the lectures.
Overall, it's good content. Java machine learning is an unfamiliar field, but there are almost no lectures introducing it, so it was very helpful. However, since I mostly use Java for web, it's a bit difficult.
Just as Java played a pivotal role in transitioning from web (jsp) to mobile (android), it is expected that Java will play a major role in data analysis (weka). Thank you for your good review.
I was very satisfied with the feedback on quick questions and the supplementary content for the missing content. The lecture content was also explained well so that even beginners could easily understand it, which was very helpful. Thank you.