CARE Driver Distractions Study
- November 21st, 2009
- in
This was a recent report requested to provide information for the Distracted Driver Summit held at UAB on December 3, 2009. A variety of driver distraction causes are discussed and compared.
This was a recent report requested to provide information for the Distracted Driver Summit held at UAB on December 3, 2009. A variety of driver distraction causes are discussed and compared.
The authors have developed a Hotspot Identification Taxonomy (HIT) that organizes the various methods for viewing hotspots. Basically they are defined as follow:
Effective use of the HIT model required four interrelated activities: data-collection, linear hotspot identification, presentation and assessment.
As part of their youth-alcohol program, the Alabama Department of Economic and Community Affairs requested a special study to focus on the development of Youth-DUI countermeasures. This report is in three sections. The first is a summary of recommended countermeasures in prioritized order based upon estimated cost-benefit. The recommendations are based upon the detailed analysis performed for Alabama and reviews of potential countermeasures given in the literature. The second and third sections of this report provide the detailed data analysis that was originally performed for the State of Alabama for CY (calendar year) 2003. These have been updated using CY 2007 data.
This paper presented the status of CARE with emphasis on its newly created GIS capabilities as of January, 2007. Emphasis is placed on the following aspects of the CARE/GIS system:
Several additional innovations have been made to CARE since early 2007, and those interested are urged to review the CARE pages within this site.
This paper reviews the use of two new metrics for the process of assessing the significance of attributes in a database when two subsets of the data are compared. Traditional statistical techniques are useful, and the sample size in public safety databases usually allows the normal approximation to the binomial distribution to be used in comparing proportionate values. For example, the comparison of the proportion of alcohol related crashes on Saturdays would show an very highly significantly higher proportion than that for non-alcohol related crashes. However the new metrics go a step further than this in that they provide a clear intuitive grasp to the user as to exactly how much more is occurring, not in terms of proportions but in terms of number of crashes (for the traffic safety example). The metric is called Maximum Gain, and it measures directly the number of crashes over and above that which is typically expected. This provides a clear indication to the user of just what the potential gain is by applying a countermeasure related to the attribute (e.g., applying selective enforcement on Saturdays). It is not realistic to think that this gain would include all of the crashes for the attribute value; rather, it is realistic to view the maximum gain to be the total over-represented amount.
This paper explores a data mining process in which the original dataset is first transformed through a variable subset selection process followed by the application of a machine learning algorithm. A variable ranking technique, called the Sum of Maximum Gain Ratio (SMGR), is applied. This technique computes a score that is based on the over-representation of attribute values. Essentially, SMGR is the ratio of the number of cases that could potentially be reduced by an effective countermeasure to the total number of cases associated with the over-represented value. SMGR was shown empirically to provide comparable results to alternative techniques, but it had significantly improved runtime performance.
This research built on the foundation of the Critical Analysis Reporting Environment (CARE), which was developed at the University of Alabama to mine crash reports submitted by investigating officers in the field. The research extended CARE capabilities by developing neural network algorithms to automatically learn potentially problematic attributes over time. The system was piloted and tested using records from Walker County, Alabama.
This paper presents an early (2003) review of CARE that was published in IEEE Computer, the flagship publication of the IEEE Computer Society. The major points made in the paper include:
This paper describes a neural network approach to automatically select crash countermeasures. The approach uses as a base the CARE IMPACT analysis that mines the crash database for the most critical attributes. However, while CARE bases its countermeasure selection on the “maximum gain” that can be obtained by eliminating attribute value over-representations, two additional algorithms were introduced to determine the most significant variables and to rank the attributes, namely: (1) a Euclidean distance approach, and (2) an ellipse distance approach. An evaluation function is constructed from the neural network learning, capturing the decision making strategy and then used continuously to select countermeasures.