JMP Wish List

What inspired this wish list request?  The SVEM prediction columns are currently being saved with an ad hoc correction applied to the the vector of predictions by fitting a simple linear regression of the SVEM predictions to the original prediction column and then saving the predicted values from this simple linear regression.   What is the improvement you would like to see?  Even if it is a hidden or scripting-only option, I think it would be nice to have an option to disable this correction and simply save the raw SVEM predictions.    Why is this idea important?  I would use this for testing the properties of the predictions with and without the correction. I have talked to some other users who would also be interested in this ability for the purposes of testing SVEM.   ... View more

What inspired this wish list request? SHAP values are invaluable for explaining ML models.   What is the improvement you would like to see? The inventor of SHAP values, Scott Lundberg, has a paper Consistent Individualized Feature Attribution for Tree Ensembles. It says "SHAP interaction values can be interpreted as the difference between the A. SHAP values for feature i when feature j is present and B. the SHAP values for feature i when feature j is absent."     Why is this idea important? My experience is that prior to modeling, a majority of collaborators ask if interactions between factors can be investigated.  Explainable ML is the key to model acceptance, and obviously ML models are deployed where nonlinearity is suspected to dominate (otherwise regression would be sufficient).     ... View more

Hi everyone I suggest for jmp staff create add-ins JMPGPT - JMP® app and class to access ChatGPT API from OpenAI™, such that JMPGPT allows you to easily access the ChatGPT API provided by OpenAI. With this add-in, maybe you can load a list of prompts for specific use cases and engage in conversations with ease. If you're new to ChatGPT and prompt engineering, the JMPGPT add-in will be a great way to learn.   Raaed ~✍ ... View more

With the increase in speed of deployment and usefulness of AI tools like ChatGPT could JMP use this as part of EasyDoE and other platforms to help users design an experiment as well as the analysis   Using a simple chat window linked to the AI it could ask you questions about the factors, constraints etc and this could then use JMP to build up the design of the experiment. Other companies are taking this approach to AI.   AI seems to be the future and integrating it into tools like JMP will be essential for the growth of data analytics in the (very near) future. Not implementing this soon could mean users look elsewhere for the simplicity of interactions with AI.     ... View more

What inspired this wish list request?  Past personal need and experience with colleagues who are implementing partial signal analysis algorithms to analyze data.     What is the improvement you would like to see?  Signal processing functions for standard results found in pythons scipy, including filters(digital & analog), FFT (spectrum), noise, convolution, correlations, transformations, peak analysis, SNR, THD, and vibration analysis. If possible I would go further and add functions for simulating signals as well, this would also be great for teaching and well as synthesizing signals for specific needs with good control and intuitively.    Why is this idea important?  I believe one of JMP's major strengths is the low entry level that it allows of it's users to be able to do complex tasks quickly while assure it's integrity and also assist users in gaining a deep understanding of processes.  I feel the area of signal processing lacks this immensely and feel that the folks at JMP can contribute greatly to this field.  This is a major field with increasing traction and I believe this would align well with JMP's renewed emphasis for targeting academia.      ... View more

Hi JMP users, I used to run some non-parametric comparisons for each pair using the Wilcoxon method in a repeated-measures designs through Analysis -> Fit Y by X report. I would like to know if it is possible to develop an effect-size calculator (add-in) for that purpose, please. Thank you. ... View more

Provide a way for users to define their own model and optimization techniques in the model screening platform, including tools from R or Python, enabling almost any model to be called by this platform, including techniques that have not yet been added to JMP, or perhaps never will be.    This could be similar to defining custom JSL functions where the user would define a function for performing the analysis and returning summary statistics and model parameters, and then another function to launch the analysis based on those model parameters returned from the initial optimization. This could be all neatly wrapped in an Associative Array along with names and descriptions. I expect this would most commonly be managed as part of an Add-in, if needed I think it could always be managed that way. The custom functions should allow users to call R or Python within their function.  In that case the function they define to launch the analysis might need to launch a new window they created to interface with R or Python. I briefly hoped this was already possible as the XGBoost Add-In for JMP Pro  add-in appears to add a new model to the platform, but it sounds like this is a special case that is hard coded in JMP itself. ... View more

In medical research the outcome under study is frequently rare, so the data are highly unbalanced. The model perfomance can significnatly improofed by correction for inbalancing by over- or undersampling. Therefore, I would suggets that all prediction model platforms should have the inbalanced classification option. Even better, the Model screening platform should have the inbalanced classification option included.   Regards,   Lu ... View more

The addition of quantile regression to the generalized regression platform was/is very welcome. It would also be nice if pseudo rsquare and other goodness of fit measures for quantile regression could be provided in the results. Thanks. ... View more

Since JMP 16, either the normal and pro versions, the menu option for cross-validation is gone! To use cross-validation with decision trees, you need to use the Model Screening platform (new in JMP Pro only), and you can't modify any tree building option in there. This is more a downgrade than an upgrade! It was very convenient to have cross-validation along with all other options in a self-contained analysis platform.... Please bring it back! ... View more

According to a Stack Overflow post, the maximum line length in SQL Server Management Studio is "65536*Network Packet Size - which by default is around 1500 bytes." When recently exporting SQL scoring code through the Formula Depot in JMP Pro for a Bootstrap Forest model the code was formatted as only three lines of incredible length. SQL Server couldn't even parse the code and simply paging down in the code became a time consuming process. When copied into Word, the document was >2,000 pages long!   This could be avoided by either A) fully parsing the code with line breaks and indentation using common best practices or B) adding line breaks every so often into the generated code at a few logical points such as at the start of each new CASE statement with a rule to force in a break before the limit is reached.  ... View more

Can we have tables of indirect effects and total effects in Structural Equation Models?  When there are mediating effects in the model, the direct regression effects table gives only a partial picture of the full model. When the indirect effect has an opposite or nullifying effect, then the total effect coefficients give the full picture of the entire model.  Bollen (1989) gives formulas for the total effects of latent variables on other latent variables and observed variables which might be useful in doing prediction (wishful thinking). ... View more

Predictive models with binary Y that allow cross validation, eg k-fold or leave one out, the confusion matrix reflect only the last iteration of the cross classification. For example in 5-fold cross validation the training set has 80% of the data and the Validation set shows 20%.  The validation set should combine all 5 folds of the data to show 100%.  For leave one out it is even more useless as the validation set only has a single observation.   ... View more

As a JMP Partner in Brazil, I teach people how to use JMP to solve problems using data, and I’ve received some feedback on the language options available within JMP. While multiple languages are supported, Portuguese is not yet one of them. Including Portuguese would be a significant step toward expanding the JMP community in Latin America. Brazil is a huge market with tremendous potential! Although there are thousands of JMP users here, not all are fluent in English or Spanish. Our competitors’ software includes Portuguese support, which often makes it the preferred choice for some users. Adding a Portuguese option would greatly benefit the JMP community and increase adoption across Brazil. Thanks! ... View more

What inspired this wish list request?  During a webinar this morning by JMP about predictive modeling for collinear and high-dimensionnal data, when using Fit Model platform and trying to save model column formula, it was a bit harder to determine which columns were leading to a formula and which ones were only saved values. In the Fit Model platform, the number of saved columns options can quickly become very large, example : In some platforms, it may not be very clear if the saved columns will contain formula (for example probabilities for SVM in the case of classification).    What is the improvement you would like to see?  In JMP, there is already a visual way to differentiate column values from column formula through the little "+" icon in the data table. By using and extending this icon to other platforms that could enable to save and export formulas, it would better highlight which saved columns contain a formula : Using and extending this icon/design would better facilitate the expectations and lead to quicker actions.   Why is this idea important?  By this simple change, it could provide a more uniform user experience and better homogeneity in design interfaces. ... View more

What inspired this wish list request:  JMP seems to have a measure for k-means to choose an optimal number of clusters.     What is the improvement you would like to see:       I believe hierarchical clustering has some statistical measures such as silhouette to statistical choose an optimal number of clusters.   If this does not currently exist within JMP, perhaps you will consider adding it.       Why is this idea important?   A statistical measure would be more accurate than the current graphical technique     ... View more

Hello, I am a physician researcher. In biology and clinical measurements, many times it is difficult to analyze data numerically because of patient heterogeneity, no real baseline, etc. What we do have is trends, i.e. "where is this (value) going". Biologically, many times the changes, trends and trajectories are more important than absolute values, or just more reliable. Therefore when I'm following up patients or results from biological samples, it is not uncommon for me to need to see what is the trend over time of a group of patients or samples. Then I want to compare them, cluster them, etc. There are some techniques to do that, very well reviewed and exemplified here: https://doi.org/10.1093/bib/bbad390 Another important aspect of this (local similarity analysis) is the ability to synchronize signals as they evolve over time.   Having these features in JMP would be invaluable!   Uriel. ... View more

What inspired this wish list request?  Documenting your models and data can be difficult but it is becoming more and more important.     What is the improvement you would like to see?  Some time ago I came across Model Cards (Google Model Cards, [1810.03993] Model Cards for Model Reporting (arxiv.org)). Something similar could be implemented in JMP and I would also implement it for datatables. Easy platform which can be utilized to define specific things related to your datatable/model (description/details, intended use,...). In best case this could be printed in some format which could then be utilized outside of JMP.   Why is this idea important?  It is important to document your models and data but it can be a hassle. Currently (in my opinion) it is very difficult to do in JMP in any sort of efficient manner which would be easy to access, fill, update and read later. ... View more

What inspired this wish list request?   Models in the Predictive Modeling platform do not have a default capability to produce confidence intervals or prediction intervals around model predictions; however, JPM has provided a way to generate confidence intervals around the predictions in the profiler using bagging (https://www.jmp.com/support/help/en/18.0/#page/jmp/example-of-confidence-intervals-using-bagging.shtml#). Unfortunately, there does not seem to be a way to produce prediction intervals in the profiler for predictive models using JMP (I am a JMP 17 user so I am unsure if this capability has been added, but based on the JMP 18 help entry I don't believe this capability has been added). I suspect it is feasible to produce these prediction intervals based on my understanding of how the confidence intervals are calculated.     What is the improvement you would like to see?    With JMP 18's renewed its focus on prediction intervals in the profiler (https://community.jmp.com/t5/JMPer-Cable/Prediction-Profiler-enhancements-in-JMP-18/ba-p/718342), my hope is that prediction intervals for predictive models will be added. It would also be helpful to generate column formulas for confidence intervals and prediction intervals for predictive models.   Why is this idea important?   Prediction intervals are useful to judge whether the data generated by the proposed model looks like the actual data, often called a "posterior predictive check" in the Bayesian literature. If the data generated by the model does not resemble the actual data, then it may be desirable to consider a different model. Please see a description of this concept, along with a simple example, from John Kruschke's book "Doing Bayesian Data Analysis" below:   "The fifth step is to check that the model, with its most credible parameter values, actually mimics the data reasonably well. This is called a posterior predictive check... One approach is to plot a summary of predicted data from the model against the actual data... The predicted weight values are summarized by vertical bars that show the range of the 95% most credible predicted weight values. The dot at the middle of each bar shows the mean of the predicted weight values. By visual inspection of the graph, we can see that the actual data appear to be well described by the predicted data. The actual data do not appear to deviate systematically from the trend or band predicted from the model. If the actual data did appear to deviate systematically from the predicted form, then we could contemplate alternative descriptive models. For example, the actual data might appear to have a nonlinear trend. In that case, we could expand the model to include nonlinear trends... We could also examine the distributional properties of the data. For example, if the data appear to have outliers relative to what is predicted by a normal distribution, we could change the model to use a heavy-tailed distribution..."         ... View more