Increasing concerns about terrorism have made border security a growing concern. As technologies have grown more sophisticated so have the ways in which those with suspicious motives can circumvent them. Technologies such as text analysis, sentiment analysis, and named entity recognition tools that use machine learning to help border security officials identify potential fraudsters are vital to keeping the nation secure and ensuring that those who try to enter the United States by deceptive means are stopped.
In 2016, the U.S. Border Patrol (USBP) apprehended 415,816 people nationwide, up from 337,117 in 2015, while arrests by U.S. Immigration and Customs Enforcement (ICE) numbered 114,434. Named entity recognition tools help border officials to identify those who present at the border. They form part of text analytics, which is a process to transform unstructured data into meaningful information and can be used beyond border security in other forms of fact based decision making.
This form of text analytics can be used by border security to find possible threats during screening of individuals at or near the border, as well as to identify things that need follow-up. It can also be used to forecast issues at borders that may be a threat in the future. Text analysis involves retrieving information, using natural language processing to mine the text, extracting the information, and data mining.
One aspect of the broader area of data mining is entity resolution, which involves finding and mentions of the same entity within and across data sets and linking those mentions together. To do this, named entity recognition tools have to deduplicate information, link records, and canonicalize it. This allow border security officials to research the people that present at the borders. Essentially the software identifies and catalogues what it determines are unique entities, which may include names, places, and dates.
For example, a border security official is processing the entry request of a tourist named James Smith. James Smith is one of the most common name combinations in the world. The border security official needs access to as much information as possible. Text analysis including identity resolution software can help him or her determine if James Smith and Jamie Smith are in fact the same person, for example, or examine his travel movements for patterns that may raise concerns.
In a broader context, government agencies can also use text analysis? sentiment analysis software to determine negative or positive attitudes in data such as social media posts. This can be helpful in identifying a person?s political views and beliefs and determining risk.
All of these tools can not only help to make a border security official?s job easier, it can also identify risks that may not immediately be apparent and compensate for things such as typos or variations in spellings that may cause red flags to be missed.
More than 76.4 million foreigners arrived at U.S borders between July 2015 and July 2016, according to the International Trade Administration. As the number of travelers arriving at U.S borders continues to increase, such tools will be crucial to maintaining the safety of our citizenry while still ensuring that those legitimate travelers presenting at border security are allowed to go safely on their way.
