Set Up the Similarity Parameter

The Similarity parameter setup depends on the input data nature and goal you want to achieve. First ask yourself the question whether the columns order in the CSV input is important or not. Said in other words, ask whether the columns order can be changed in any manner without changing the data sense. For example, if the columns identify the hours of the day, days of the month or weeks of the year, then it is to expect that the columns order is important. On the other hand, if the columns identify such data as the height, mass and age, then their order should not be important.

– If the columns order is important, then choose the A. Shape – in parts or C. Proximity – in parts setting. Please remember that – in parts – means the similarity within the context of the curve neighboring points, which form parts of the curves.

– If the columns order is not important, then choose the B. Shape – in points or D. Proximity – in points setting. Please remind that – in points – means the similarity within the individual points of the curves regardless the context of the neighboring points of its parts.

The second question related to the Similarity parameter setup depends on your intended goals. You either want curves with approximately identical values or curves with similar shape in the resulting clusters. In other words, you either want the curves in the diagram to be closer or their shape is important.

– If the curves shape is important, then select the A. Shape – in parts or B. Shape – in points setup. Please remind that – shape – means the curves shape similarity regardless to the curve position in the diagram.

– If the curves distance is important, then select the C. Proximity – in parts or D. Proximity – in points setup. Please note that – proximity – means similarity of the values of the curves, i.e. their proximity in the diagram.