Please help!!! I am new to Weka, and have to do a project for school regarding classification. Is there anyone familiar with Weka that would be willing to point me in the right direction? Thank you!
I attached the data file for the project. We have to use Weka to build a model that could be used to predict lateness.
Are you allowed to use any type of model you want? Are you looking to create a model for Late vs Not or Degrees of Late vs Not?
Something you could do is create groups, Not late, late 1-5 mins, late 6-10mins, etc, then run a cart model or random forest.
WEKA tool has several applications, and among them, "Explorer" is probably the most popular one.
In "Explorer" application, the data set to be analyzed must be imported in "Preprocess" tab at first. The csv-formatted files are available, but I recommend you to convert it into arff-formatted file.
In order to apply the classification techniques such as naive Bayesian classifier, decision tree and nearest neighbor, the data set MUST HAVE a class, a dependent variable in categorical form.
Maybe, the last attribute "lateMins" is the dependent (target) variable, but it is numeric. So, you have to "discretize" the "lateMins" attribute so that a class for classification analysis will be obtained.
In "Preprocess" tab, click the [choose] button and you can find discretize filter for discretization.
If you have imported your data set in "Preprocess" tab, the [Classify] tab will be enabled. [Classify] tab provides many classification algorithms, and you can see the list of them by clicking [choose] button. After an algorithm is selected, clicking the [run] button will provide you the result of classification analysis.
- Badges
- Science topic