[ad_1]
The information that the police organizations store on detentions or criminal incidents, as well as the notices to the police departments, have an enormous value to solve future cases that can be raised.
Analyzing manually this large amount of data en busca de patrones puede llevar mucho tiempo y sus results pueden ser limitados. We propose to demonstrate how you can help the most advanced methods of text analysis to improve results. In concreto, queríamos descubrir textuales y geoespaciales processables relacionados con la lucha contra la trata de seres humanos (Figure 1) and otros delitos.
Para ello, teníamos que pensar en como mejorar el processo con la technología. In concreto, se trataba de procurir capacitas de las que se beneficiarían las personas que trabaja día día en la policia, en lugar de un analista o un científico de datos. In the last instance, what we wanted was to improve the effectiveness of the police investigators through the use of text analysis to relieve incidents related to the traffic of persons and other criminal patrons, and we continue to do so. informs the visual. Fortunately, los métodos de analysis de textos que hemos aplicaido en otros proyectos y que categorizan automáticamente los data y buscan tendencias, entidades (personas, lugares, objetos) da connexions entre fuellianany documents.
This flow of work and focus can be seen in Figure 2, which details the process and the analysis applied to political information. The political text documents were processed in the visual flow of text analysis of the graphic and assisted interface, where natural language processing and pre-defined text analysis techniques were applied, which included the solution, extraction of entities, classification, perfilado de los data de texto y más. This approach based on visual flows resulted in standardized tables and lists for analysis in Visual Analytics with the objective of exploring, investigating and visualizing the results of our analysis. This process is proportionally enormously accelerated and the time and results in the information retrieval terminal involved in a series of policy investigations. This process identifies patrons of robo, violence and trata de human trivia from 45,000 files with 45,000 title data.
Gran parte de nuestros results market and regulation that will use SAS visual text analytics (VTA) are important to identify and are precursors to patrons delictivo. Se Utilizó un conjunto de Reglas Conceptuales and Integración de Codigo Abierto for Extracter, geocodificar y categorizar las ubiaciones por tipo. At the same time, it is a written rule that is additionally directed towards calls. This rule used a combination of street numbers, street words (Avenue, Street, Drive, etc.), district indicators (N, S, E, W) and filler words that represented the literal street name. . This is a module, Pudimo filter los incidentes ocurridos junto a centros escolares, como se muestra en la figure 3.
Tras Extracter los nombres completos de las calles, se pasaron por un processo de Python (geopia is used) que produjo una latitud y una longitud para cada reción. Las Coordenadas results in geocoding a la inversa. This was done to recover the direction from the newly discovered coordinates and obtain a more detailed direction.
Creating geocoding and reverse geocoding directions:
- Original name: 920 SAS Campus Drive Cary, NC 27513
- Geocoordinates: 35.815658, -78.749284
- Geocoding inversion: SAS Global Education Center, 920 SAS Campus Drive Cary, NC 27513
As it has been seen in the previous example, when performing the inverse geo-coding, you can obtain additional information such as the hotel, the gas station, the school and other key names of this region. This is the permission of additional information agrupari los lugares additional and the creation of one taxonomy VTA que classification los lugares por tipo. These are 10 types of product creams, which include gasoline, restaurants, hotels and colleges, among other products. Cuando se combina con análisis adicionales, is to set up additional categories and the corresponding new campus structures to make the points relevant and enter for the analysis of visual analytics. This additional entry point allows for exploratory analysis and rapid discovery of relevant and interesting facts. An example is the location of a robbery with a firearm occurring in front of a primary school. We were able to geolocalize and classify unstructured information based on time, place and event type by means of geo coding, location evaluation and weapon extraction. analysts.
Se desarrollaron additional rules in VTA for extracting the vehicles from the documentation that the police stored. This rule used a combination of characteristics of a vehicle, such as color, brand, model, year, type and key descriptors of a vehicle. When observing the combinations of these features, we extract many vehicles from the information we have and provide additional and useful information to deepen and observe the patterns in the collection from the collection and provide additional and useful information to deepen and observe los patrones en la collection de documentation política. In figure 4, examples of identified vehicles are shown.
Many logical concepts that contain a line on the red diagram (Figure 5). The blue nodes are documents of origin, the yellow ones are directions and the orange ones are mentions of arms. This visualization allows for 40,000 document policy matches, trends, and possible methods. Muchos de las relaciones y las coincidencias seria imposible de detecter mediate una revision manual sin la ayuda de la extraction de las concepts y las visualizationaciones. En la Figura 5 se pueden ver numeros ejemplos de tendencias potential interesting interesting. Regarding the 2005 Chevy blanc, about posting information about Podemos. Otro ejemplo es el examen de la frequencia y las tendencias con las que se hace referencia a armas o directciones specificas en los reports.
Las Reglas relacionados con la trata de seres humanos se desarrollaron utilizando IA y metodos estadísticos en SAS Visual Analytics for identified patrons and torno and conceptual concepts. In figure 6, the highlighted term is similar to “prostitución” en el conjunto de información, identificamos immediatente terminos relacionados con la trata, como “albergar”, “reclutar” y, in particular, “menor denunciante”.
From here, using IA methods and additional rules related to threats, coercion, blackmail and escapes, we could point out incidents that put human trafficking in relief in Figure 7) risego como physical violence against women/adolescentes who could be related to human trafficking directly or could create a trafficking situation in the future.
Juntando todo esto, pudimos utilizador los métodos geospaciales mentioned previously para obtener aquellos incidents que implican trata de seres humanos or un risego de trata de seres humanos to put them at the disposal of the investigation of the investigation comouradou 8. que un investigator o un agent de policía podría utilizar
In summary, our objective was to show how, as part of minimal structured data, we took advantage of text analysis capabilities to identify patterns in information that could be used intuitively. Aunque los departamentos de policía disponen de metadata adicionales relacionadas con estos incidents, es posible que dichos metadata only permitan identify un delito primario, como un incidente de consumo de drogas, mientras que en la text informaciónadicuuest. Mo el riesgo de trata de seres humanos. Adequate, application methods are similar to textual texts of otras fuentes de dates for investigation, como las pista or transcriptions declarations or lamada, asi ayudar and filtering, with classification and usage policies to understand information.
More information
It’s a contenido in tradition and bass and in a blog post written by Kirk Swilly and Tom Szabo.
[ad_2]
Source link
