• Pre-sales Consultant
  • 400-6240-800

Product Overview

DEEPFINDER, PERCENT’s intelligent security analysis system, is a “big data + AI” system platform that safeguards national public security; it mainly includes three parts: big data management platform, data cognition engine and data application service. It can provide public security organs and government law enforcement departments with strong technical and tool support for efficient and comprehensive integration, analysis and use of data and mining of inherent value from massive data resources to the maximum.

Product Functions

Efficient and Agile Data Governance and Consolidation

Big Data Management Platform (BD-OS) is used for standardized cleaning, full integration and unified storage management of massive multi-source heterogeneous data accessed by structured, semi-structured, unstructured and online user’s behaviors. Through data access, transformation, exploration, modeling, experiment and use, the whole life cycle management of the model can be realized, experimental domain and production domain of the model connected, and the huge gap between fields separately managed by decision makers and developers bridged.

Flexible Data Modeling

By putting top experts’ knowledge in business algorithm model through machine intelligence, the modeling of any different fields can be realized, a plurality of different modeling modes can be applied to the same field, and data model can be dynamically expanded and adjusted. Four categories, such as, security for stability maintenance, case investigation, commanding & scheduling and social management are offered in addition to more than 110 kinds of application models. It supports analyst to conduct one-click application, provides visual presentation of relationship among data, and flexibly adjusts models according to the change of business morphology.

Panoramic Search and Full-text Retrieval

A one-stop search service system based on public security’s all-dimension data resource combines static data (population data, case data and criminal data, etc.), which are scattered in various fields and systems, with real-time dynamic data (social media data, electronic fence data and internet of things, etc.) to support multi-dimensional search methods, such as, classification, association and fuzzy, etc. Dynamic knowledge graph is used to build a powerful retrieval engine, which supports the retrieval and analysis of unstructured data, such as, text and pictures. At the same time, it supports detailed granularity analysis of structured data, and automatic data association of search results.



Multidimensional Correlation Analysis

It gathers the information of people, object, space & time, organization and virtual identification information, constructs space-time matrix analysis that combines time analysis, spatial analysis, statistical analysis and line expansion analysis, offers perspective picture of the relationship among entities in three dimensions, and displays it visually.


Visualization and Thematic Analysis

By using natural language processing and speech recognition technology, it can realize intelligent data extraction, analysis and visual display; problems of data can be analyzed and detected independently, and one-stop data management of data analysis can be carried out. In terms of application in fields like government, and criminal investigation of public security, it supports to analyze a special event in the form of special topic, which conducts targeted analysis on the development trend of a special event from various dimensions, such as, time and space. The analysis results can be viewed and exported in the form of reports.

Product Functions

Efficient and Agile Data Governance and Consolidation

Big Data Management Platform (BD-OS) is used for standardized cleaning, full integration and unified storage management of massive multi-source heterogeneous data accessed by structured, semi-structured, unstructured and online user’s behaviors. Through data access, transformation, exploration, modeling, experiment and use, the whole life cycle management of the model can be realized, experimental domain and production domain of the model connected, and the huge gap between fields separately managed by decision makers and developers bridged.

Flexible Data Modeling

By putting top experts’ knowledge in business algorithm model through machine intelligence, the modeling of any different fields can be realized, a plurality of different modeling modes can be applied to the same field, and data model can be dynamically expanded and adjusted. Four categories, such as, security for stability maintenance, case investigation, commanding & scheduling and social management are offered in addition to more than 110 kinds of application models. It supports analyst to conduct one-click application, provides visual presentation of relationship among data, and flexibly adjusts models according to the change of business morphology.

Panoramic Search and Full-text Retrieval

A one-stop search service system based on public security’s all-dimension data resource combines static data (population data, case data and criminal data, etc.), which are scattered in various fields and systems, with real-time dynamic data (social media data, electronic fence data and internet of things, etc.) to support multi-dimensional search methods, such as, classification, association and fuzzy, etc. Dynamic knowledge graph is used to build a powerful retrieval engine, which supports the retrieval and analysis of unstructured data, such as, text and pictures. At the same time, it supports detailed granularity analysis of structured data, and automatic data association of search results.



Multidimensional Correlation Analysis

It gathers the information of people, object, space & time, organization and virtual identification information, constructs space-time matrix analysis that combines time analysis, spatial analysis, statistical analysis and line expansion analysis, offers perspective picture of the relationship among entities in three dimensions, and displays it visually.


Visualization and Thematic Analysis

By using natural language processing and speech recognition technology, it can realize intelligent data extraction, analysis and visual display; problems of data can be analyzed and detected independently, and one-stop data management of data analysis can be carried out. In terms of application in fields like government, and criminal investigation of public security, it supports to analyze a special event in the form of special topic, which conducts targeted analysis on the development trend of a special event from various dimensions, such as, time and space. The analysis results can be viewed and exported in the form of reports.

Product Functions

Efficient and Agile Data Governance and Consolidation

Big Data Management Platform (BD-OS) is used for standardized cleaning, full integration and unified storage management of massive multi-source heterogeneous data accessed by structured, semi-structured, unstructured and online user’s behaviors. Through data access, transformation, exploration, modeling, experiment and use, the whole life cycle management of the model can be realized, experimental domain and production domain of the model connected, and the huge gap between fields separately managed by decision makers and developers bridged.

Flexible Data Modeling

By putting top experts’ knowledge in business algorithm model through machine intelligence, the modeling of any different fields can be realized, a plurality of different modeling modes can be applied to the same field, and data model can be dynamically expanded and adjusted. Four categories, such as, security for stability maintenance, case investigation, commanding & scheduling and social management are offered in addition to more than 110 kinds of application models. It supports analyst to conduct one-click application, provides visual presentation of relationship among data, and flexibly adjusts models according to the change of business morphology.

Panoramic Search and Full-text Retrieval

A one-stop search service system based on public security’s all-dimension data resource combines static data (population data, case data and criminal data, etc.), which are scattered in various fields and systems, with real-time dynamic data (social media data, electronic fence data and internet of things, etc.) to support multi-dimensional search methods, such as, classification, association and fuzzy, etc. Dynamic knowledge graph is used to build a powerful retrieval engine, which supports the retrieval and analysis of unstructured data, such as, text and pictures. At the same time, it supports detailed granularity analysis of structured data, and automatic data association of search results.



Multidimensional Correlation Analysis

It gathers the information of people, object, space & time, organization and virtual identification information, constructs space-time matrix analysis that combines time analysis, spatial analysis, statistical analysis and line expansion analysis, offers perspective picture of the relationship among entities in three dimensions, and displays it visually.


Visualization and Thematic Analysis

By using natural language processing and speech recognition technology, it can realize intelligent data extraction, analysis and visual display; problems of data can be analyzed and detected independently, and one-stop data management of data analysis can be carried out. In terms of application in fields like government, and criminal investigation of public security, it supports to analyze a special event in the form of special topic, which conducts targeted analysis on the development trend of a special event from various dimensions, such as, time and space. The analysis results can be viewed and exported in the form of reports.

Product Advantages

Dynamic Knowledge Graph

Dynamic knowledge graph is a leading knowledge extraction & fusion technology developed and innovated by PERCENT. Through machine intelligence plus human brain wisdom, people, objects, organizations, time & space, and virtual identification in the real world can be mapped into digital world, and the relationship between them is automatically constructed to support users to carry out analysis and intelligent decision- making.

Dynamic knowledge graph stands at the core of connecting the real world and the data world.

• Dynamic Data Fusion

Flexible Data Modeling

• Real-time Knowledge Evolution

Natural Language Processing

PERCENT’s Natural Language Processing (NLP) uses machine learning and in-depth learning technology to realize part-of-speech tagging of word segmentation , named entity, sentiments analysis and text classification. Among them, the NLP technology can achieve 98.97 % accuracy in word segmentation recognition, and 91.45 % accuracy in entity recognition, leading the industry. It can quickly extract knowledge from open documents on the internet, construct enterprise-related entities and mapping relationships, realize semantic-based in-depth understanding and knowledge-based information, and support machine to make intelligent decision.

• Natural language processing based on in-depth learning can significantly improve the effect of recognition.

• It supports multilingual natural language processing.

• Natural language processing based on knowledge graph enhances the accuracy of semantic understanding.


Dynamic Knowledge Graph

Dynamic knowledge graph is a leading knowledge extraction & fusion technology developed and innovated by PERCENT. Through machine intelligence plus human brain wisdom, people, objects, organizations, time & space, and virtual identification in the real world can be mapped into digital world, and the relationship between them is automatically constructed to support users to carry out analysis and intelligent decision- making.

Dynamic knowledge graph stands at the core of connecting the real world and the data world.

• Dynamic Data Fusion

Flexible Data Modeling

• Real-time Knowledge Evolution

Natural Language Processing

PERCENT’s Natural Language Processing (NLP) uses machine learning and in-depth learning technology to realize part-of-speech tagging of word segmentation , named entity, sentiments analysis and text classification. Among them, the NLP technology can achieve 98.97 % accuracy in word segmentation recognition, and 91.45 % accuracy in entity recognition, leading the industry. It can quickly extract knowledge from open documents on the internet, construct enterprise-related entities and mapping relationships, realize semantic-based in-depth understanding and knowledge-based information, and support machine to make intelligent decision.

• Natural language processing based on in-depth learning can significantly improve the effect of recognition.

• It supports multilingual natural language processing.

• Natural language processing based on knowledge graph enhances the accuracy of semantic understanding.


Dynamic Knowledge Graph

Dynamic knowledge graph is a leading knowledge extraction & fusion technology developed and innovated by PERCENT. Through machine intelligence plus human brain wisdom, people, objects, organizations, time & space, and virtual identification in the real world can be mapped into digital world, and the relationship between them is automatically constructed to support users to carry out analysis and intelligent decision- making.

Dynamic knowledge graph stands at the core of connecting the real world and the data world.

• Dynamic Data Fusion

Flexible Data Modeling

• Real-time Knowledge Evolution

Natural Language Processing

PERCENT’s Natural Language Processing (NLP) uses machine learning and in-depth learning technology to realize part-of-speech tagging of word segmentation , named entity, sentiments analysis and text classification. Among them, the NLP technology can achieve 98.97 % accuracy in word segmentation recognition, and 91.45 % accuracy in entity recognition, leading the industry. It can quickly extract knowledge from open documents on the internet, construct enterprise-related entities and mapping relationships, realize semantic-based in-depth understanding and knowledge-based information, and support machine to make intelligent decision.

• Natural language processing based on in-depth learning can significantly improve the effect of recognition.

• It supports multilingual natural language processing.

• Natural language processing based on knowledge graph enhances the accuracy of semantic understanding.


Application Scenario

The system combs through information of cases, conducts deep mining, comprehensively analyzes person, objects and organizations relevant to cases and other associated relations, combines it with dimensions like time and space to carry out multi-dimensional statistical filtering so as to realize comprehensive analysis, research and judgment of cases.

With this system, case investigation no longer has “no clue” as it uses known clues to intelligently guide reconnaissance. After case information, case type and on-site inspection information are typed in, the system will automatically classify, then provide steps to guide the investigation according to different case types and clues. This greatly reduces the workload of civil policemen as they no longer need to consult various legal instruments, case guide and other information.

It quickly expands to find people and events associated with suspects, and constructs social relationship graph of suspects by combining time and space correlation analysis and multi-dimensional statistical filtering. At the same time, analysis, research and judgement is carried out to comb through the information on organization structure of criminal gangs, identify potential core leaders, intermediaries and other gang members, and detect behavior characteristics, activity trajectory and abnormal behavior types of criminal gangs.

It monitors the latest trends through dynamic analysis of physical tracks and online tracks of specific people or groups, investigates other data, such as, calls among them, locations and banks after gathering clues are detected, then verifies criminal clues, predicts potential behaviors of criminals, prevents repeated crimes and thus greatly reduces crime rate.

It supports multidimensional analysis and mining of business data and social data, combines online public opinion, public data on internet, forums, micro blogs and others with personnel information at public security department , then conducts multiple relationship analysis, behavior analysis, trajectory analysis, so as to achieve active prediction , early warning and prevention of major events.

Based on established regional security assessment model, it analyzes crime rate, efficiency of investigating and solving cases, and the amount of police resources in the region, calculates based on three indicators, namely, destructive power, control power and safeguard power, forms regional public security risk index, visually display comprehensive nationwide public security situation, realizes optimal allocation of police resources, and sends out warning on regional public security situation.

It aggregates bank account information, fund transfer information, ATM withdrawal information, multimedia image video and others, identifies abnormal transaction behavior in complex fund transaction network through multi-level transaction flow analysis and space-time analysis of a large number of accounts, determines suspicious accounts, and detects suspects by combining geographic information and image video information of ATM withdrawal.

It collects hot topics and hot events discussed by netizens on news media, forums, blogs and other websites, and comprehensively displays popularity and content of related topics. In addition, by using analysis on hot topics discussed by online public and relations, it identifies development trend and spread path of message, sends out early warning on trends, helps public security sector to put information of online public opinion under surveillance.



Integrated Analysis of Cases
Case-guided Detection

The system combs through information of cases, conducts deep mining, comprehensively analyzes person, objects and organizations relevant to cases and other associated relations, combines it with dimensions like time and space to carry out multi-dimensional statistical filtering so as to realize comprehensive analysis, research and judgment of cases.

With this system, case investigation no longer has “no clue” as it uses known clues to intelligently guide reconnaissance. After case information, case type and on-site inspection information are typed in, the system will automatically classify, then provide steps to guide the investigation according to different case types and clues. This greatly reduces the workload of civil policemen as they no longer need to consult various legal instruments, case guide and other information.

Excavation of Criminal Gang
Criminal Acts Analysis

It quickly expands to find people and events associated with suspects, and constructs social relationship graph of suspects by combining time and space correlation analysis and multi-dimensional statistical filtering. At the same time, analysis, research and judgement is carried out to comb through the information on organization structure of criminal gangs, identify potential core leaders, intermediaries and other gang members, and detect behavior characteristics, activity trajectory and abnormal behavior types of criminal gangs.

It monitors the latest trends through dynamic analysis of physical tracks and online tracks of specific people or groups, investigates other data, such as, calls among them, locations and banks after gathering clues are detected, then verifies criminal clues, predicts potential behaviors of criminals, prevents repeated crimes and thus greatly reduces crime rate.

Major Events Early Warning
Security Situation Analysis

It supports multidimensional analysis and mining of business data and social data, combines online public opinion, public data on internet, forums, micro blogs and others with personnel information at public security department , then conducts multiple relationship analysis, behavior analysis, trajectory analysis, so as to achieve active prediction , early warning and prevention of major events.

Based on established regional security assessment model, it analyzes crime rate, efficiency of investigating and solving cases, and the amount of police resources in the region, calculates based on three indicators, namely, destructive power, control power and safeguard power, forms regional public security risk index, visually display comprehensive nationwide public security situation, realizes optimal allocation of police resources, and sends out warning on regional public security situation.

Capital Flow Analysis
Online Public Opinion Monitoring and Early Warning

It aggregates bank account information, fund transfer information, ATM withdrawal information, multimedia image video and others, identifies abnormal transaction behavior in complex fund transaction network through multi-level transaction flow analysis and space-time analysis of a large number of accounts, determines suspicious accounts, and detects suspects by combining geographic information and image video information of ATM withdrawal.

It collects hot topics and hot events discussed by netizens on news media, forums, blogs and other websites, and comprehensively displays popularity and content of related topics. In addition, by using analysis on hot topics discussed by online public and relations, it identifies development trend and spread path of message, sends out early warning on trends, helps public security sector to put information of online public opinion under surveillance.



Inquiry

Name
Email
Mobile Number
Company Name
Inquiry Description

* Please provide correct mobile number for us to get back to you

Submit