OptiSegment provide explainable results and supports explainable artificial intelligence (XAI) philosphy.
OptiSegment can work with only a dataset with flexible number of columns. Optionally, below parameters can be set for the segmentation:
- Number of clusters desired: can be an exact number of min-max number of clusters, so OptiSegment will find the best number of segments automatically. (only for OptiSegment v.4.0, not included in OptiSegment RFM version)
- Optional Datasets: Originally OptiSegment can work with 1 single dataset but details can be provided for the OptiSegment trilogy (please refer to https://www.optisegment.com/rfm for further details and some use cases, and please note, all data sets except relation is optional).
- Length of Journey: Optionally, a parameter for the joruney of subject can be provided. OptiSegment can create journey (like in Customer Journey Analysis) from the changes of segments and the Length parameter for the updaes is optional. For example, if you want to see the customer journey with 1 month, than optisegment will create a customer journey graph for segment changes in the last 1 month.
- Current Segments: Optionally, you can provide the current segments and OptiSegment checks for the specific data points, where optisegment decides the data points are mis-segmented.
After creating the segments from the data set, OptiSegment creates below reports:
Report 1: Assignment of each data point to a segment. This is the classical segmentation with segments for each data point.
Report 2: Borderlines of each segment for each parameter in the data set. The data set might have multiple columns (parameters, features) and OptiSegment creates the borders for each segment / parameter combination.
Report 3: Outliers in the data set. If current segments are provided, OptiSegment reports the outliered data points from the current segmentation as well.
Report 4: Data Set Matches. If optional datasets provided for the details of subject and object in OptiSegment trilogy, than the data matching will also be covered by OptiSegment.
Report 5: Numerical achèvements for each segment. The term achievement can be varying for the segmentation problems. For example, profit, net incode, number of items sold, stock turnover rate, number of registered students, number of movies sold etc. might be different for the problem. OptiSegment finds the achivement of the segment success by the numerical columns provided on the Relation data set of OptiSegment Trilogy.
Report 6: Input -> Achievement relation. OptiSegment can create a connection between each input and its effect to the achievement in the Report #5.
Report 7: Multi Segmentation comparison. OptiSegment can compare the behaviour of each data point from different segmentation algorithms. With its advanced technology, OptiSegment can hold multiple state-of-art segmentation algorithms and compare their results and finds the data points where all algorithms agree or disagree.
Report 8: Visual representation of data set, data points and segments in 3D screens.