User's Guide
xSDR is designed in order to improve functionality and usability. The graphical user interface provides enough information for the user to perform a large number of actions. There is also a large number of informative messages that direct the user while using the program.
Also there is a full detailed user's guide available for everyone that wants to look deeper in the program and take advantage of every feature. In the following paragraphs we will give a small guide for the absolute beginner to the Dimensionality Reduction.
The tool is divided in five layers. The first layer is the data input. Below you can see the main screen of the program.
In this stage you will select your data source. You can select between SQL server, MySql, Oracle DB or a text file. Once you select the desired type of data source you press create and the program guides you to the creation of the new dataset. Alternatively you can press fetch and your stored datasets (if any) will appear. Then you can load the one you want by pressing "Load" or you can re-configure it by double-clicking on it. Then you can proceed to the next step.
The second stage is the data management layer. In this step you can manage your dataset. You can replace any string values with numeric, perform mathematical operations or normalize some attributes. Below you see the main screen of the data management stage.
The message you can see is a common warning that informs you that there are some non - numeric values in your dataset that need to be transformed before you proceed to the Dimensionality Reduction.
The third stage is the Dimensionality Reduction step which comprises the tool's cornerstone. Below you can see the dimensionality reduction screen.
This step is very easy to use. You select the type of the algorithm you want to execute (Distributed or Central) then you configure the dimension you want to reduce your dataset to and any other parameter (if any). Then you just press execute and the Reduction is done! If anything goes wrong you will receive an error message.
The fourth step is to visualize your results. For this action we use Microsoft's MS Charts library. Below you can see the data visualization screen.
Usage of this tab is straightforward. There is no much configuration during this step and all you can do is to select every possible diagram or compared attributes before and after the dimensionality reduction.
The last layer is the interaction with the weka tool. As we can read from weka's original web site: "Weka is a collection of machine learning algorithms for data mining tasks. [...] Weka contains tools for data pre-processing, classification, regression, clustering, association rules, and visualization."
Weka is a powerful and high configurable data mining tool, but its usage is out of this guide's scope. We encourage you to take a deeper look at this project and move xSDR's functionality even further!
Info:
This is only a quick start guide. Only a few features of the tool are demonstrated. We strongly recommend to download the detailed user's guide and take advantage of every feature. Below you can find links for the Greek and the English version.
You can also contact the development team. We will be happy to help you and try to answer any question you may have.