This site uses different types of cookies, including analytics and functional cookies (its own and from other sites). To change your cookie settings or find out more, click here. If you continue browsing our website, you accept these cookies.
I want to learn how to do a predictive model for the reduction in size of something. At the end what i want to be able to do is put in people details and then it would use the model learnt from the data to see how it thinks they will do.
So for this exercise is to say for Tumor reduction (i know this is not how tumors work)
I have attached some fake data i have created for 10,000 patients. 50% of them have a Type called Lung and 50% have it called skin.
How would you go about doing this? I have never used predictive models like this before so if im way off the mark let me know.
I am just looking for some advice or even if your aware of a good guide that i can use id really appreciate this.
If you are looking for a guide to get started with predictive modeling, the Data Science Learning Path on community provides an overview of the process, some Data Investigation techniques, and a quick overview of the algorithms available in Designer.
In the dataset you uploaded, you have some covariates (type, Smoker, Age, Gender) but you also have time series data (week over week). The best type of model will depend on what type of prediction you are planning on making. If the "new" data you use to make a prediction relies on the covariates, then standard predictive tools (brown tools in Designer) will work. If you will have the size from the first few weeks, then you may lean towards a Time Series model (orange tools in Designer).
Do you want to predict the change overall tumor size from week 1 to week 12? Or do you want to predict the size at every week?
Do you have any other possible data for each ID?
At this point you have very few other possible features associated with this dataset, so building a valid model might be difficult. You really only have Smoker, Age, and Gender, as I'm guessing you'd want separate models for each type of tumor.
It would be every week i want to predict the size of the tumor.
I am not very worried about the accuracy of the model from this limited information as i know it doesnt have anywhere near the level of detail to allow you do produce a valid prediction, i am just looking to understand the process of predictive modeling in alteryx and how i could do this.