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.
Recently I am working with some predictive analysis and I found the description of Naive Bayes quite fit my requirement
Creates a binomial or multinomial probabilistic classification model of the relationship between a set of predictor variables and a categorical target variable.
However, when I started to use it, the target variable can only be string format while I would like to predict a number (price range to be precise). Is that the limitation of the model or what I can do to configure it? Plus for these classification models such as random forest, boosted mdoel, etc. is there a way to find the confidence interval with score tool? If not, how can I create the interval?
The Naive Bayes Classifier tool is set up to only use string's as the target variable due to most classes of categories being in that format. Also, in some rough testing adjusting the macro to allow non string types to be the target variable, this looks to be a limiation in the R used in the macro itself. There is however a quick work around for this. You can place a select tool before the Naive Bayes Classifier tool and change your price range field from a numeric type to a string field type. This will allow you to select that field in the Naive Bayes Classifier tool.
For your second question, yes. You can calculate a confidence intervale with the Score tool. From the Score tool's help menu: Include a prediction confidence interval:If this option is checked, confidence intervals will be calculated using the specified confidence level. If scoring with a model created by ORE, the original estimation table must exist in the database in order to calculate confidence intervals.