To find terms which have a significant effect on the modelling of the data, Synthace builds two models for every term: one which contains the term and one which does not contain the term. These models are compared to see if the presence of the term has a significant effect on how well the model fits the data (to learn more about how this works read here).

How the identified significant effects relate to the response are shown on the Find Effects tab as a set of interactive plots. They allow you to assess and scrutinise whether the identified significant effects actually look reasonable, or whether effects which were not found to be significant by the approach might still be beneficial to the model fit. Based on this, you might choose to remove or include an effect in the selected set.

Where they are present, two factor interaction effects will also be plotted to help you understand the behaviour of your system. These are not the same as a full model fit, which is representative of all the effects, but gives you an idea of the direction and magnitude of the effects and their interactions.

The main effects plots will show your chosen subset of response or transform data plotted with the fitted model for that specific effect term.

Plots will be created appropriately to represent categorical, mutually exclusive (left) or numerical (right) factor types. Whilst a continuous model is plotted across the categorical groups in the effects plots, this is just to help guide the user in understanding trends between the independent groups and these are not modelled continuously during model fitting.

Two factor interaction plots demonstrate the impact one factor has on the output of another. One factor is plotted against your chosen response with the second factor used to colour all data points and model fits in the third dimension.

Interacting with the sliders beneath two factor interaction plots will alter the transparency to highlight just those data points in the selected range. This way it is easy to see the effect of one factor when it’s set points are high or low, for example, with regards to the other factor.

If you want to visualize the effects plot for two factor interactions with a different factor used as the x-axis, you can switch the factors around by clicking Invert Interaction Terms.

Clicking Invert Interaction Terms, swaps the terms around in the graphs, so the x-axis factor becomes represented by the colour scale and vice versa. Clicking on the colour keys will show and hide that specific data to aid clarity in interpreting the plots.

Well done for making it to the end of this tutorial.

To learn about other modelling techniques, click here.