On this page you can make a quiz used to determine a particular outcome. For example, the outcome could be based on the question, "What superhero are you?", and the questions can be anything ranging from your favorite color to what kind of music you like. All you need to do is create a title, questions, at least two outcomes, and possibly some images and descriptions.
Once you create the quiz, you can share it with anyone, and with user feedback, the application will use AI to learn how to predict results.
Note: When you submit an image created for this quiz, you grant me non-exclusive permission to use, modify, and distribute the image.
Logistic regression uses a weighted sum of responses to determine the outcome. Random forest takes the average of many decision trees, which may work better for more complicated relationships. By default, the first few results are random, and then logistic regression is used until enough responses are gathered for a random forest model.
What is the largest number of responses to be used for training at full weight? Keep blank for no limit.
Setting this value can allow you to mostly train a quiz before sending it out (so the results are fairly consistent).
Maximum weight of remaining responses. Set to 0 or keep blank to only use at most the number of responses above.
For example, if you set the largest number of responses to be training at full weight to be 50, and this value to be 2, then if there were 250 responses, the first 50 would have a full effect, and the next 200 would have only twice the influence on the model, instead of 4 times the influence if there were no limit.