nxcurve 0.6.1

draws RNS,QNX and BNX curves and their auc

QualityCurve

Dimensionality reduction ( DR ) is a data transformation process which provides a low-dimensional ( attribute or variable ) representation of high dimensional data sets. This with the following purposes: noise reduction, storage space reduction, data visualization, efficient data processing and to concentrate the important information in fewer variables than the original set. A performance visual measure in DM is topology preservation. Quality curves RNX, proposed by Lee and Verleysen, evaluates performance generating a graphical representation of topology preservation. Nowadays there is a tool for topology conservation evaluation of DM algorithms, developed also by Lee and Verleysen (2009) but this tool is implemented only in Matlab. Here a problem arises because Matlab is limited and cannot be implemented in other technologies. here, we are going to implement, in python, a software evaluation module of curves RNX in order to be used in other technologies.

New Features!

• RNX curve and area under the curve
• QNX curve and area under the curve
• BNX curve and area under the curve

Development

• Grahp for the corranking matrix
• LCMC from a coranking matrix (local continuity meta criterion)
• Error Handling

Want to contribute? Great!

License

MIT

Free Software, Hell Yeah!