Parameters

Similarity - video 11, video 12, video 13 and video 14

By the mean of this parameter setup, the user defines the evaluation method for the curves similarity at the generation of the clusters. The differences between the parameter possible setups are crucial and simple:

shape- means the similarity of the curve shapes regardless to the curve position in the diagram

proximity- means the similarity of the curve corresponding values, i.e. their proximity in the diagram is evaluated

in parts- means the similarity within the context of the neighboring points of the curves, which form parts of the curves

in points- means the similarity in the individual points of the curves regardless to the context of the neighboring points of the parts of the curves.

It is to comment that the similarity of the curves within the context of the neighboring points (in parts) means that the order of columns (the curve points) is important since the curve parts are defined just by the neighboring points, which form parts of the curves. Vice versa, the similarity of the curves in the individual points of the curves (in point) means that the order of columns (the curve points) is not important. Therefore, the order of the columns (curve points) can be changed freely regardless to the result.

For more information about Similarity feature please follow SimplexManual.

Modification - video 15 and video 16

While setting up this parameter, the user defines the method of the input data modification. At the real use of the perSimplex product, it is always necessary to prepare the input data in a manner allowing reaching the sufficiently specific shape of the diagram curves in order to generate the clusters. In many cases, the original data are ready for use directly and without any modification. In such cases, the unchanged original data are used for the generation of the clusters; this corresponds to the Without data modification setting.

Usually, it is necessary to change the original data by the mean of suitable setup of the Modification parameter in function of the user's goal related to the generation of the clusters. By the mean of use of the individual settings, the input data values are modified to values between 0 and 100% and the modified data curves are subsequently used for the generation of the clusters.

The setups of the Modification and Similarity parameters can be combined arbitrarily or as needed.

For more information about Modification feature please follow SimplexManual.

Analysis- video 6

While setting this parameter, the user defines whether the statistical data distribution should be accepted for the generation of the clusters or not. Usually, the real data have values, which are not distributed homogeneously across the entire range. Usually, certain portion of data values show anomalies when compared with most of the data, whose values tend to get closer to one or several dominating values.

Due to either A. Statistical analysis - in row or B. Statistical analysis - in columns setting, the application regards the real statistical data distribution at the generation of the clusters so that it is more sensible for the evaluation of numeric values similarity in the intervals with higher data density.

For more information about Analysis feature please follow SimplexManual.

Stress - video 17

The Stress parameter can be used for A. Shape - in parts and C. Proximity - in parts settings (Similarity parameter). Using the Stress parameter allows the emphasis on certain part of the curve while generating the clusters. Said in other words, the similarity of the curves becomes defined mainly by the mean of their similarity in the stressed part.

If the Stress parameter is equal to 0, then no part of the curve is stressed. This means that the shape or proximity of curves is of equal importance for its entire length.

If the Stress parameter is set up to any value between 1 and 100, then you will emphasize certain part of the curve. The parameter value shows the location of this stressed part. In case of value equal to 1, the curve start is stressed (diagram left side) and in case of value equal to 100, it is the curve end (diagram right side). Thus, 50 corresponds to the curve center etc.

For more information about Stress feature please follow SimplexManual.

Strictness - video 8 and video 10

The Strictness parameter may acquire values from 0 to 99. As the parameter value increases, the mutual similarity of the curves in clusters increases as well, and, at the same time, the number of identified division clusters increases too. Thus, for higher values of the Strictness parameter, we will obtain more clusters with curves with higher degree of mutual similarity.

The Strictness parameter can be used for any setup of the remaining parameters of the SimplexDivide application.

For more information about Stress feature please follow SimplexManual.