Skip to content
PREDICTING DRUG AND ALCOHOLABUSE TREATMENT COMPLETION BY LOOKING AT GUN VIOLENCE
Share
Explore

The Treatment Score (cont.)

Processing/Feature Engineering
In order to "score" an area, I needed to transform my TEDS-D data so that each row represents one CBSA area and I needed to be able to count the total number of clients for the different labels in each column for that area.
To do this, I one-hot-encoded each column, grouped the data by the CBSA codes, then summed the resulting columns. This then allowed me to treat each label in the original dataset as a variable to use in my score functions.
The Long List of Variables
The table below describes all of the variables I am using to generate this score.
I included the sign of each coefficient to distinguish between "good" and "bad" variables.
"Good" variables have a positive impact on the score.
"Bad" variables have a negative impact on the score.
image.png
image.png
To shorten some notation,
image.png
Some Score Functions

image.png
Some Score Plots

Discharge Scores (y) vs. Referral Scores (x)
image.png
Length-Of-Stay Scores (y) vs. Referral Scores (x)
image.png
Length-Of-Stay Scores (y) vs. Discharge Scores (x)
image.png
The Everything Scores
image.png
Want to print your doc?
This is not the way.
Try clicking the ⋯ next to your doc name or using a keyboard shortcut (
CtrlP
) instead.