Solving serial crimes involves investigating thousands of leads, items of evidence and other bits of information. So many pieces of data can quickly overwhelm the ability of the human brain to analyze to arrive at a conclusion. Very labor intensive, this effort often takes months and even years. Meanwhile, the serial criminal remains unknown and free to commit more crimes.
Using computers to solve serial crimes is not new. While investigating a number of savage killings in Washington state in the 1970s, Robert Keppel laboriously punched the pieces of information onto IBM computer cards (many younger types probably don't even remember punch cards) so the computer could correlate them. The computer analyzed all the data, albeit pretty slowly, and came up with a likely suspect: Ted Bundy.

7,000 homicides and nearly 6,000 sexual assaults, the HITS database is exceeded only by those of the FBI and the Royal Canadian Mounted Police in the detail of information available on violent crimes.
The Pacific Northwest National Laboratory (PNNL) in Richland, Wash., is currently developing Computer Aided Tracking and Characterization of Homicides, or CATCH. Working closely with the Washington State Attorney General's HITS unit, the objective is a set of advanced innovative data visualization and analysis tools for much more efficient analysis of large crime databases like HITS. The PNNL is a U.S. Department of Energy National Laboratory operated by Battelle; this particular research is being funded by a grant from the National Institute of Justice. The project is part of PNNL's Law Enforcement Initiative.
CATCH assesses likely characteristics of unknown offenders by relating a specific crime case to other cases and by providing a tool for clustering similar cases that may be attributed to the same offenders. To accomplish this, CATCH has several tools-including clustering maps, query tools, geographic maps, timelines-each of which is designed to present a different view of the case data, with emphasis on a particular property. The tools let the user remove cases that are deemed unrelated and add cases from the database that are considered similar to the cases being analyzed.
At the heart of CATCH is Artificial Neural Network (ANN) technology. A form of artificial intelligence, neural networks are computer systems designed to find hidden patterns in complex data like the HITS database.
ANNs handle such information much like the human brain would, but they do so much faster and more efficiently. The CATCH's ANNs learn to cluster similar cases from approximately 4,000 murders and 3,000 sexual assaults (some solved, some unsolved) residing in the HITS database. CATCH groups together those crimes that share similar features such as modus operandi; the manner of death, type of weapon used, locations; signature characteristics of the offender (including sex rituals), mementos taken from the crime scene and other parameters describing the victim and offender.
The result is an easy-to-read ANN graphic presentation on the CATCH screen showing the distribution of crimes over a two-dimensional map, with darker boundaries separating areas of similar crimes. Any group of crimes clustered together could be the work of one person.
Additionally, CATCH can suggest a possible profile of a suspect by comparing unsolved cases with similar ones that have already been solved. The bottom line is, CATCH can help link cases in ways human investigators would not necessarily see. Plus, there is no way a human can look at 7,000 cases and compare over 200 variables.
How well does CATCH work? CATCH has already helped to remove one unsolved murder from the HITS database-although no killer has been caught. CATCH determined that two separate sets of chopped-up body parts were from the same victim, rather than from two individuals as originally thought. According to PNNL project manager Lars Kangas, CATCH may prove more useful for catching serial rapists than serial murderers. That is because there is usually a description of the rapist. In contrast, serial murderers tend to get caught by accident rather than from investigations that put the pieces together to identify a particular subject.
Sophisticated computer-based systems like CATCH are the wave of the future in criminal investigations. With area murder investigative systems being developed, each with substantial databases, these systems will be needed to make them valuable tools in solving crimes and putting criminals behind bars.
One important capability is to be able to compare violent crime cases from different jurisdictions so investigators can approach the investigation knowing that similar cases exist. PNNL has turned over the CATCH system to the Washington state's Attorney General's Office for further evaluation. It is also currently looking for more funding to modify and upgrade the first prototype of the system.