Overlapping Nodes in CHAID (Decision Tree) in SPSS Modeler
I occasionally encounter nodes in CHAID models (in SPSS Modeler) that seem to have overlapping values, such as:
Above, the split is on a continuous variable ("Fulfillment in: Working at a job..." etc. is based on a Likert-scaled item). I'm unclear about how to interpret the nodes---for example, Node 4 is <= 5.000 but Node 5 is 5.000,6.000. I notice that there are brackets but don't know what they represent.
Or is this because I've configured the build-options incorrectly? They are currently set to:
- Build new model
- Build single tree (Generate model)
- Exhaustive CHAID
- Max tree depth set to custom value: 10
- Default stopping rules (percentages), Default Ensemble rules, and all Advanced settings also are the defaults
Thank you in advance for any guidance.
There aren't any overlaps. SPSS uses the same notation for ranges of values (a.k.a. "intervals") as you might find used in a calculus course. The rounded parentheses indicate that the interval does not include the endpoint, while square brackets indicate that the endpoint lies within the interval.
So, the middle node of the tree is marked "(6, 7]" but, since the variable values are integers, it's really just cases with a value of 7 that fall into that node. For a Likert-scaled item such as this one, you may want to tell SPSS to treat the variable as ordinal, rather than continuous.
- Reading .sav file into R
- UnicodeDecodeError: 'charmap' codec can't decode byte 0x81 in position 55: character maps to <undefined>
- Get the sub colnames of a dataframe from SPSS to R (read.spss)
- ODBC error when trying to connect MS Excel to SPSS file on fileserver
- Why does converting a nested python dictionary into a pandas dataframe result in "has no attribute 'items' error?
- Simple contrasts output from SPSS in R
- SPSS If statement with multiple conditions
- Importing SPSS dataset into Python
- SPSS Custom Tables Nesting
- Change value labels or delete specific value labels
- pyreadstat read and write spss without data loss
- Efficient way for converting Yes/No to 1/0 in Excel or SPSS
- Fill in missing values in a row based on another column SPSS syntax?
- Pivot variables in SPSS with `CASESTOVARS` and specify the resulting column names
- Convert string YYYY-MM-DD to MM/DD/YYYY date in SPSS?
- R - changing values to labels permanently in labelled data
- SPSS match files results in empty column
- calculation with different values in SPSS
- How do I shorten super long variable labels of an SPSS file in R?
- Dynamically remove columns with no variance in syntax
- SPSS Syntax: spliting one categorical variable into multiple yes/no variables
- Translating a simple data check method from SAS into SPSS (crosstabs for 3+ variables fed into output dataset)
- SPSS: How to access values of variables, where the names of the variables I want are the values of another variable
- How do I write a DO IF command asking to compute something ONLY if the value of that case is missing? [SPSS]
- Creating one variable from three, SPSS
- How to properly invoke Python 3 script from SPSS syntax window using SCRIPT command (+ additional problems during runtime)
- What's a pythonic way (native function in pandas) to count occurrences of a certain value within cases (SPSS COUNT equivalent)?
- In a dataset where individuals have multiple rows of data per enrollment, how can I make a variable that indicates enrollment instance (1st/2nd/etc.)?
- Import .spv file into R / convert into .csv
- SAS no longer automatically converts user-defined formats to SPSS value labels