Abstract
All of the material to this point would be at best academic if modern algorithmic risk forecasts were unable to inform practice. This chapter turns briefly to porting new forecasting procedures to the settings in which they will be used. There are technical issues, but often the major obstacles are interpersonal and organizational. Implementation can be the most challenging and time consuming step that must be anticipated as the risk algorithm is being developed.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
- 2.
There is no credible, technical way to attach such labels. A good strategy is to discuss the nature of algorithmic reliability with stakeholders and use their label definitions unless they are statistically misinformed.
- 3.
The discussion purposely has been centered on R. Increasingly, however, effective and appropriate machine learning algorithms for risk assessments can found in Python environments. Most would argue that Python is a far better general purpose language than R, but for statistical applications R will likely be a better choice. R has been build specifically for data analysis and related tasks. Python has not. For very large datasets, both R and Python now call procedures, often written in more efficient languages such as Java or C++, that can dramatically reduce processing time.
- 4.
This kind of error is common. For a variable called “Employment,” an individual recording the data might just leave the entry blank if the individual is not employed.
References
Berk, R. A. (2016) Statistical Learning from a Regression Perspective second edition New York: Springer.
Berk, R. A., Barnes, G., Ahlman, L. & Kurtz, E. (2010) When a second best is good enough: a comparison between a true experiment and a regression discontinuity quasi-experiment. journal of Experimental Criminology 6(2) 217–236.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Berk, R. (2019). Implementation. In: Machine Learning Risk Assessments in Criminal Justice Settings. Springer, Cham. https://doi.org/10.1007/978-3-030-02272-3_8
Download citation
DOI: https://doi.org/10.1007/978-3-030-02272-3_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-02271-6
Online ISBN: 978-3-030-02272-3
eBook Packages: Computer ScienceComputer Science (R0)