Do we always need to bin the continuous variables or is BayesiaLab able to directly handle these variables?
Quote 0 0
Continuous variables always have to be binned, i.e. discretized. BayesiaLab offers a broad set of tools for discretizing continuous variables. These tools are available within the Data Import Wizard, in the Node Editor and in the context of the Re-Discretization function. [*:31rbn988]Decision Tree (supervised discretization when you have a target variable)[/*:m:31rbn988][*:31rbn988]Density Approximation (for fitting the continuous density function)[/*:m:31rbn988][*:31rbn988]K-Means (another way to try fitting the continuous density function)[/*:m:31rbn988][*:31rbn988]Equal Distances [/*:m:31rbn988][*:31rbn988]Normalized Equal Distances (to prevent the negative impact of outliers)[/*:m:31rbn988][*:31rbn988]Equal Frequencies[/*:m:31rbn988][*:31rbn988]Manual Discretization: to allow you to set bins directly on the distribution or density functions[/*:m:31rbn988]Let's take the example of the continuous variable C104 that is selected in the screen shot above.Selecting the Manual Discretization shows you the distribution function. Clicking on the Switch View button displays the density function. The default threshold, indicated by the blue vertical line, is set to the mean value of the variable.Manual Discretization:Thresholds can be added and removed by right-clicking on the graph, and can be modified by selecting them. The automatic discretization algorithms can be either selected by using the Type button or in the Manual mode by clicking on Generate a Discretization. [*:31rbn988]Type: the selected discretization algorithm will only be applied at the end of the importation process[/*:m:31rbn988][*:31rbn988]Generate a Discretization: the discretization algorithm is run and the computed thresholds are displayed in the graph.[/*:m:31rbn988]
Quote 0 0
I have a dataset, which contains a weight for each sample. The weight value varies between 1 and 600. In Bayesialab's weighting functionality there is possibility to use the weights as they are or to normalize them.What is the difference between them? At least I did not see any differences in my model.
Quote 0 0
Dan
When you choose to normalize the weights, their values are changed in order to have the total sum of the weights equal to the number of observations in your dataset. If not normalized, the weights are kept unchanged. If you have an observation with an associated weight of 600, this is thus equivalent to duplicating this observation 600 times.This is thus quite strange to have the same result. You can double check the active option by hovering hover the database icon that is in the lower right corner of the graph window.
Quote 0 0