Model Fitting Using the "Random Search" and "Randomly Iterated Search" Method
An alternative method is using a combination of the Random Search and Randomly Iterated Search methods. The process is as follows:
- Begin by clicking the Manually Reduce or Manually Expand button to set a general regression model.
- Click the Random Search button and wait as the system finds several different models. If the search appears to stall, press the ESC button to stop it.
- Select the best model, then click the Randomly Iterated Search button to search the model in detail. You may repeat this process several times.
- If any equations are marked red, click the Auto Reduce button to remove them. Repeat the reduction process, then click the Randomly Iterated Search button again to find the best model.
You can use the Manually Reduce/Expand or Auto Reduce/Expand buttons in any order to improve, freely expand or reduce the model. It is strongly recommended to use the Randomly Iterated Search option after every expansion and reduction operation.
New models will be added to the collection after following these steps.
You can improve, expand, or reduce any previous model from this collection at any time.
It is strongly recommended to fit the first model again, as the heuristic technique used by ndCurveMaster for curve fitting improves and enables the discovery of better models, even when repeated. This is because:
- The discovery method for the best models will be different each time,
- And the models may also be different.
Therefore, it is advised to fit the models several times from the first linear model, creating and adding new, better models to the collection. Finally, you can choose the best model.
Another option is to use the Random Search and Randomly Iterated Search methods separately.