Analysis: 0. level: Overview
Basing on the following list the user can choose the needed modules directly. If You can not interpret yet, how the provided modules should be used, please, look for the advising services.
In case of each module an OAM is required. This OAM can not have any title for the row- and column-heads. Each X-value should be transformed to integer (rank) and each Y-value should be also integer (rank or nominal). The free services work only under such circumstances, if the volume of the matrix do not exceed the threshold of 20*20. A more complex job can be run having appropriate access code. If You have a huge job, please, do not hesitate, to write an email (s. above) for further information.
For each running phase it is necessary to define a job-code (max. 12 vaules and the 1. number can not be the zero). If a job-code should be unfortunately already busy, a system message will be sent immediately.
There are 3 steps to solve a job (OAM):
- First the initial values should be set. Hereby the numbers of rows (objects) and X-columns (X-attributes), the keywords and the shifting (if the Y-values were transformed to the positive range) have to be defined. Please, pay attention: the Y-column should be not included to the parameter setting.
- Second an anonym OAM should be uploaded (having only integer values, consequently without any head-component of rows and columns).
- Third the well-prepared job can be calculated and the results will be provided immediately by the screen.
Special effects: The offline similarity analysis can support the handling of primer data (not only ranking values), or the interpretation of models without Y (cf. monoton Y-vector) as simulation of universal similarity, or the building of models with multiplicative (and other: e.g. sin) components.
Each online calculation is based now on the lp_solve (5.5.0.7).
