• 

• 

• 

Amazing flowers, aren’t they ? Let’s imagine us walking across a wildlife area and observing beautiful flowers we’ve never seen before. We like them so much and want to see them again but all we know is only that they look like Iris. Checking contents of our backpack we realize that we’re lucky to have a ruler with us. So, we can measure some characteristics:

• sepal length
• sepal width
• petal length
• petal width

We’re getting closer to classification problem. The main idea is to identify the exact type of Iris by a a set of collected characteristics. So, the problem definition is clear. Scientific explanations of classification are not supposed to be covered by this post. There are lots of classification algorithms. To take meaningful decision about choosing appropriate one scientific background is desirable. Implementing existing classification algorithms ourselves is like implementing “wheel”, there is no time to implement things which have already been implemented and work quite well. As a result, there are libraries with collections of machine learning algorithms.

One of such libraries is Weka. It will be presented in this post to solve Iris classification problem. Iris data set is often used in classification tutorials due to its simplicity.

If you’re not a newbie in solving classification problems and have used Weka before, you may have a question: “What new information will I find in this post?”. It’s indeed not the first Weka-based classification tutorial, so to make this information unique I’ll present in the next parts of this article own wrapper library that uses Weka for solving classification problems.

Don’t miss the next parts, they will have technical details, code samples and interesting results.