Machine Learning using Boosting Regression in JASP free software | Supervised learning

[@haroldasraz]

Thank you for an excellent presentation. JASP has become extremely impressive. The team and community behind Jamovi and JASP are making a statistically significant software of a gem.

[@haroldasraz]

Can you specify which subjects the model should use? Let's say I trained my model and collected approximately 20 subjects, and now I want to see how well the model predicts disease for this sample whilst being trained on the prior data set.

[@TravelerTriumphs]

I am moving from SPSS to JASP, and will be using this for my quantitative analysis callses to my MA sociology students, found your presentation really useful.

[@xiaoyulan8662]

Really amazing presentation and clear interpretation! Thank you. is there any chance of sharing the ppt slides you employed in the presentation?

[@ZZU04]

Sir i have watched your video and totally agreed with you that there is no rule of thumb.. According to my littlr knowledge what i usually intrepret as a statistician is i compare the MAE OR MSE sets of different input combination and look minimal value.

[@uzairbaig8372]

How can we use prediction algorithm for boosting regression in latest JASP version

[@hemantz1001]

Can you upload a video on random forest ML technique on JASP platform

[@buraktiras93]

Really good presentation, thank you for that! My question is, let's say we completed the training and testing part and saw that we have a model that has high accuracy. How can we deploy it to make further predictions with new inputs?

[@VahidAryadoust]

Models cannot be deployed from JASP, as far as I know. It is an area for improvement, which you can bring up with the software developer.

[@haroldasraz]

Once you have trained a model. How can you test it on new test data?

[@panpanyang8938]

Quite brilliant!

[@黃威綸-w9p]

Really useful, but how can we apply the model into new data prediction even we do not have the correct answer. thanks your time.

[@VahidAryadoust]

Sorry could you rephrase the question? I have difficulty understanding it.

[@黃威綸-w9p]

@@VahidAryadoust Sorry for unclear question. For example, we use dataset A generate a machine learning model B in JASP. Is it possible to use model B to do prediction in other dataset while we do not have correct answer.

[@VahidAryadoust]

You should feed the new dataset as your left-out (testing) data. Simply create a new variable in the data (call it test-train). Test = 1 and train =2. Replace the test data with your new data and run the analysis again.

[@uzairbaig8372]

I did not understand. can u please share a short clip. It will be very much helpful for me. Thanks

[@Niculuzzu]

Spettacolare

[@fernandojosearaujodasilva1918]

Brilliant

[@kamilbaszczynski8649]

If I may, I have a question. Do continious predictors need to be on similar standardized or normalized scale in order to compare their relative influence on the model?

[@VahidAryadoust]

I suggest you standardize the variables if they are not on the same scale. For example, standardize them on mean = 0 and SD = 1, or other ways.

[@kamilbaszczynski8649]

@@VahidAryadoust Thank You for the answer and what about bi- and multi-nomial qualitative data? If I use variables measurer on different types of scales. You would also recommend unification? So for example min-max conversion of all variables?