NIP - An Imperfection Processor to Data Mining datasets

Every day there are more techniques that can work with low quality data. As a result, issues related to data quality have become more crucial and have consumed a majority of the time and budget of data mining projects. One problem for researchers is the lack of low quality data in order to test their techniques with this data type. Also, as far as we know, there is no software tool focused on the create/manage low quality datasets which treats, in the widest possible way, the low quality data and helps us to create repositories with low quality datasets for testing and comparison of data mining techniques and algorithms. For this reason, we present in this paper a software tool which can create/manage low quality datasets. Among other things, the tool can transform a dataset by adding low quality data, removing and replacing any data, constructing a fuzzy partition of the attributes, etc. It also allows different input/output formats of the dataset.