Big data
Intelligent technology platform capable of integrating big data analytics in real time
Big Data technology
Saltlux Technology’s Big Data platform not only provides semantic search/analysis and enterprise/public big data intellectualization but also performs advanced analytical functions.
Big data engine
Since its founding, Saltlux Inc. has established a systematic roadmap for big data technological development through insightful prior predictions. Saltlux Inc. has developed various smart data products and platforms with extraordinary performance and functionality, combining big data-based machine learning, deep learning, and knowledge graph-based reasoning.
Based on the value chain of the data ecosystem (data collection, storage, distribution, and utilization), Saltlux’s Big Data Engine can not only analyze traditional data but also solve problems at the expert level, make data-driven decisions, accumulate knowledge, and create economic values.
Why choose saltlux technology's big data
Salltux’s big data engines can effectively handle the data that traditional methods struggle with, in terms of size, variety, speed, and value.
Korea's first intergration of AI and big data technology
Builtin Verified in-memory Stream analysis
Built-in commercial engine's Powerful features
Real-time expandability and safety guaranteed
Easy intergrated operation management system provision
Intergrated Connection Support of HADOOP Ecosystem
big data engines
Discover the typical engines of Saltlux with Big data application
Collection Engine Tornado
It is a strong big data processing engine that could perform a real-time automatic parallel collection of users’ preferences from big data that are generated in various industries such as deep web, SNS, shopping sites, IoT, and streaming data. TORNADO provides an optimized big data collection environment for real-time analysis of social big data, competitors, markets and products, risk management, and customer voice recognition. Main features:- Built-in features for collecting various big data
- RSS collection
- Scenario-based collection
- Open API-based collection
- Deep web collection
- Metasearch collection
- Collection engine operation management
Semantic Search Engine Discovery
Intelligent semantic search engine (Discovery) is an AI-based search engine that understands words and documents using machine learning and deep learning. It provides enhanced search performance and search results in users' preferences by learning the data feature through data collecting and applying a machine learning algorithm that conducts Feature Learning. It also includes the built-in Deep Search function that can automatically search for the meaning of the input keyword. Main features:- Meaning (semantic)-based integrated search feature
- Facet filler and powerful question processing feature
- search by topic feature
- Popular, related words search automatic recommendation feature
- Automatic completion of search word feature
- Information (document) automatic summary and document object search feature
- Search for similar documents feature
Text Mining Engine TMS
By comprehending characteristics, meanings, associations of large-scale internal and external non-structured data, Text Mining Engine (TMS) performs semantic-based retrieval, information rearrangement, and multi-dimensional analysis. It offers non-structured data analysis functions to discover and add value to hidden knowledge. Accordingly, users can derive decisions such as knowledge utilization, customer management, risk management, research, and development, which helps you save time while searching, analyzing, and utilizing knowledge information. Main features:- Built-in powerful text mining feature
- Built-in machine learning and deep learning-based high-quality morpheme analyzer
- Built-in machine learning and deep learning-based high-performance sentence structure analyzer
- Built-in machine learning and deep learning-based object name recognizer
- Built-in machine learning and deep learning-based reputation (emotion) extractor
- Built-in high-quality hybrid-type information extractor
- Best existing built-in machine learning-based big data automatic information (document) classifier
- Automatic built-in information (document) assembler
Stream analysis engine BlueBolt
BlueBolt is a real-time streaming data analysis engine that converges and analyzes various informal human data sources and real-time machine data, such as equipment, production line logs, and sensor data. BlueBolt enables users to create optimized systems, predict production line problems, and detect abnormalities in security through real-time in-memory analysis of stream big data and complex event processing (CEP). Main features:- Strong workbench and real-time data extraction functions
- Secure data quality and analysis performance through powerful stream data query processing
- Real-time important events detection and inducing appropriate processing
- Real-time decision making support through real-time analysis and prediction
Recognition Analysis Engine CAS
By analyzing large volumes of big data with machine learning and deep learning, the cognitive engine can quickly find characteristics, meanings, and associations among data that humans have difficulty finding. Furthermore, it provides complex system analysis of ultra-capacity data, fusion analysis between voice and text, and fusion analysis between image and text. Main features:- Object name recognition analysis feature
- Emotional recognition analysis feature
- Knowledge/Social Network Analysis feature
- Voice recognition+text integration analysis feature
Visual Analysis Engine Rainbow
Rainbow helps detect hidden patterns, predict, and assess the future by visualizing big data and the analysis result from various perspectives. Not only does it visualize simple individual data, but it also converges, reorganizes, and provides a visual analysis of data through dynamic dashboards. Main features:- Supports real-time connection of various data source
- Select to extract and combine real-time data
- Various interactive visual analysis
- HTML5-based web publishing (report)
BIG DATA CASE STUDIES
See how our customers have been utilizing Big Data technology
Intelligent spatial data was built by fusing structured and non-structured big data and text data based on spatial data such as administrative and private information. Thereby, users can use various analysis models and workflow-based analysis tools while analyzing models in the web environment. The tool also provides a multi-purpose spatial big data based analysis platform along with the service environment. Intelligent spatial big data and smart administration framework is built to proactively respond to national outstanding problems and establish future strategies.
Project Information
- Build integrated DB: The results of integrated DBs are shown in the form of a dashboard through the integrated data service that allows users to easily understand and access the spatial data.
- Data Service Development: You can download related data via a query using the data service that classifies datasets into fusion and basic data. It provides a screen division-type visualization service in which you can effectively utilize the time-series information of convergence DB.
- Build an analysis platform and develop visualization service: Spatial big data analysis platform is built based on spatial information and spatial big data. It also offers a visualization service in which spatial Hadoop-based analysis results are visualized and displayed on a map.
- Social-spatial analysis service development: The service performs emotion analysis on non-structured data by sentence unit, combines with location information, identifies relationships with issues, analyzes major topics and vocabulary trends according to regions, and visualizes them on a map.
Application technology and solutions
1. Real-time spatial big data analysis to implement smart data based on structured and non-structured spatial big data and next-generation analysis platforms.
2. HDP 3.2 packaged by Horton Works connects with Hadoop EcoSystem to build an optimal spatial big data analysis platform.
Main accomplishments
1. Received an MLIT Minister award among the building projects under the auspice of MLIT.
2. Support government-wide use of the system by establishing a spatial big data system that uses spatial information to implement scientific administration and provide customized services.
3. Establish the groundwork for common spatial big data utilization, so that central administrative offices, local governments, and public corporations can establish rational and objective policies in the public sector
- Preemptive response to socioeconomic issues in real estate, education, welfare, crime, disasters, etc.
- Support future national strategies building and decision making through scientific spatial analysis technology
4. Promote the spatial informational industry, including new industry and job creation, through the development of various utilization service models using spatial big data.
5. Spatial big data system provides users with an environment to develop analysis models easier and faster in various kinds of templates and web environment.
6. Allow users to develop new analysis models and derive analysis results by combining data and analysis functions according to their purpose.
7. Use spatial Hadoop-based location information to introduce visualization, achieve time series informatization through screen division, and visually present analysis results, such as past information and future prediction.
To establish an early risk detection system for future technology and business uncertainty, we combine and integrate the two following pieces of knowledge. One is knowledge accumulated in Samsung Electronics and Samsung Semiconductor's internal KMS. The other is the materials supplied through MOU contract with KISTI: a large number of technical documents, including overseas academic data, information about domestic conferences, various research reports, information on trend analysis of overseas science and technology, domestic and international IT news, IT professional review/blog, technical magazines, etc.
By establishing an automatic system for collecting and analyzing new technology information, early sensing capabilities are improved on new technologies and applications. Through various analysis and prediction features of new technology sensing, such as response to risks posed by the limitations of continuous technology and the emergence of non-continuous technology, venture investment to complement business risks, and establishment of M & A strategies, we have laid the foundation for internal data-based decision support.
Project Information
- Building of data collection system: Build an integrated DB for new technology sensing and predictive analysis by establishing a collection system for various internal and external types of structured and unstructured data.
- Building of data modeling and knowledge base: Various data models were built to extract various relevant information from external technical documents and social data (news, blogs, etc.) and internal knowledge and secure new technology's sensing quality. As a result, technical and corporate information knowledge bases (person/paper/patent association, technology/company/person association, company/investment institution/person association) were established.
- Building of New Technology Sensing Platform and Develop Analysis Service: Samsung Electronics built a new technology sensing/prediction/analysis platform to collect, store, process, analyze, and visualize internal knowledge and externally collected data. The tool is capable of correlation analysis, knowledge trend analysis, knowledge information network analysis, and real-time new technology sensing/prediction analysis in large amounts of data. It can also visualize and present the results of various analysis features.
Technology application and solutions
Apply Saltlux's Big data technology that enables real-time/batch analysis to identify intelligent new technology sensing and predictive analytics platform based on internal knowledge information and external technical data.
Validity of product selection
Proven excellence in product performance and quality tests through PoC
- After verifying the usability of the big data analysis platform for building a new technology sensing platform for Samsung Electronics KMS, POC is conducted three times to select the product with the best quality and performance.
- We verify all Korean and English data types in each processing stage (collection, storage, analysis, visualization, system performance) and select Big Data Suite as our intelligent new technology sensing and predictive analysis platform
Main accomplishments
1. The simple search function only yields a series of simple descriptions or individual keywords. These limited functions make it difficult to observe the technology network or identify trends in technical analysis. However, thanks to the new technology sensing system, users can easily obtain information trends, application technology, or technology of interest. Therefore, the tool meets exploratory needs of technical understanding, especially large and individual technical problems in Samsung Electronics.
2. This has dramatically increased the number of KMS connections and users from three to five times since the opening of the current system, thereby developing Samsung Electronics' knowledge management activities.
The aim is to revolutionize journalism services and develop them into differentiated government/national services by building a high value-added knowledge base from media contents, which is built by decades of domestic media organizations. We build a platform to systematically analyze large amounts of news data and analytical data for intelligent analysis. By establishing various analysis services and management systems, media organizations producing high-quality journalism information such as media-based intelligent analysis, integration, and prediction services continue to derive high-quality media contents in the future and lay the foundation for innovation in distribution support.
Project Information
- Building of news big data DB: The largest article DB was built from domestic/overseas newspapers, economic newspapers, TV broadcasting news, internet newspapers, English newspapers, regional weekly newspapers, old newspapers to the pre-90s ones.
- Building of news big data analysis platform: To analyze news big data, we build a platform for deep analysis of news big data. It includes intelligent natural language processing engines that can analyze English and Korean, machine learning-based unstructured text analysis engines, and semantic search engines.
- News big data analysis service development: The tool provides Big Kinds search for basic searches and Big Kinds professional services for professionals with semantic-based news search features.
- Through news analysis, the tool generates relevant information between news and stock index and predictive analysis. It analyzes articles, extracts citations, and offers news source analysis (network) for sources of citations. In addition, it provides a trend report feature, along with various analysis features, to users, such as deep analysis of overseas news, analysis of political news of each country, and trend analysis on national press reports, related analysis, and network analysis.
Technology application and solutions
1. Apply Big Data to analyze one-stop real-time media content big data, supporting the implementation of the unstructured news big data analysis platform based on domestic/overseas news content
2. OpenSource Apach SPARK, STORM, and KAFKA are applied to the news big data platform to create a news big data analysis platform with optimal quality.
Main accomplishments
1. By building and analyzing large-scale news data, it contributes to nationally historical asset accumulation, social change prediction, policy-making, and business opportunities.
2. In terms of economy, the tool also lays the foundation for creative economic brains to predict and respond to future situations or circumstances.
3. Lay the foundation for an innovative system where high-quality media organizations and news agencies can continue to generate and distribute next-generation media contents in the future.
4. As a leading example in the global news analysis service area, it provides a global news analysis system service and a service that analyzes differences in national media coverage on specific issues.
5. Contribute to raising social awareness of the news utilization value and create a corresponding profit model based on media content.
6. Deliver customized news and related information by analyzing news big data characteristics and dispositions of individual users and institutions.
As the number of customers using digital channels and the demand for intelligent services is expected to increase, it is necessary to improve customer satisfaction and operational efficiency by expanding AI-based services. As a result, we have established an AI big data system call center to upgrade and improve existing systems.
Real-time voice processing, language analysis processing, and large-scale distributed environment are applied to the big data analysis system that records, stores, and analyzes daily telephone consultations from Nonghyup Bank Customer Happiness Center. The tool develops a structure that enables real-time TA (Text Analysis) analysis services and the consultative data analysis results in a timely and relevant manner. Therefore, Saltlux's laid the foundation for intelligent VOC analysis and TA analysis to analyze from various purposes and perspectives.
Project Information
- Improve data storage and operational environment: Collected counseling big data is stored in the message queue, indexed through dispersion applications, stored in a big data repository, and then turned into services through text analysis.
- Build Intelligent Knowledge Management System: The knowledge information built/managed through the new KMS (Intelligent Knowledge Management System) is managed as the knowledge base at the same time. A structure that automatically distributes it to the Q & A system is developed.
- TA service composition: Provide issue cloud and issue trends analysis features- Provide emotion analysis on issue keyword.- Provide navigation feature for related topic analysis- Provide automatical clustering of consultation results by topic
Technology application and solutions
1. Big Data applied technologies includes consultation data collection class, message class, big data indexing class, utilization class, etc.
2. OpenSource Apach SPARK, STORM, and KAFKA are applied to the news big data platform to create an intelligent financial big data analysis platform with optimal quality.
Validity of product selection
1. Guarantee the integrity of Big Data Suite's large data collection and prevents any loss in real-time data transmission and reception. In addition, through distributed processing environment and in-memory analysis, the tool ensures the performance of the big data system in both large-scale aggregation and analysis processing.
2. Select products that meet the needs of an analytics system to accommodate both real-time and batch analysis of ever-increasing consultation data.
3. Select products that perform high-performance big data analysis through in-memory aggregation, statistical analysis, real-time indexing, and distributed processing.
Main accomplishments
1. Enable stable service even during peak customer consulting hours, when the system may be overloaded due to the increasing number of users. The system infrastructure can flexibly buffer the continuous growth of data or services in the future.
2. Upgrade TA analysis feature and service to improve real-time issue analysis and consultation quality evaluation
- Improve the UI and provide analytics visualization services such as search, statistics, trend analysis, etc. for counseling issues. This eventually helps extend in-depth analysis according to each topic.
- Give users the best experience when using the product through monitoring the counselor's customer consultation in real-time, automating part of the consultation quality evaluation, timely detecting and improving false consultation or response.
3. The intelligent recognition analysis feature can convert voice data into text, analyze the unstructured text, present the results, and process large quantities of data in real-time.
IT systems should be able to efficiently support the GIH process and strategic decision-making in the affiliates. The purpose is to establish a POSCO Family Integrated Information Management Network for integrated domestic and international external information from POSCO and family companies in GIH, which are analyzed and processed by experts in each field.
Intelligent analysis of structured and unstructured data was performed for GIH internal information and external information. The semantic search of related topics and issues keywords maximizes accurate information retrieval and monitoring capability. Saltlux also conducts a three-dimensional analysis of accumulated unstructured data, such as trend analysis and related topic analysis.
Project Information
- Establish an intelligent semantic search system: Extract relevant information to the search terms by analyzing in real-time the top 100 important documents, including the user's query or selected keywords (topics). As a result, we're able to build an intelligent search system retrieved in the form of a radial tree.
- Intelligent search feature development: - Clustering search results: Deliver clustering results by topic through real-time analysis of search result documents for user query- Popular search words: Statistically extract the search term history of users to display visually by ranking- Automatic classification search: Search data through automatic information (document) classification- Similar document search: Search for similar documents from various sources and show the search results in real-time
- Develop intelligent trend analysis function and service: Analyze CEO’s messages and information data, extract key keywords by period, extract issue keywords in each period, and provide visual analysis to users in various chart formats.
Technology application and solutions
Apply Big Data-based storage/search engine (DISCOVERY) and unstructured analysis engine (TMS) to perform intelligent retrieval and data analysis of structured and unstructured data for GIH internal and external information.
Validity of product selection
The DISCOVERY analysis engine and unstructured analysis engine (TMS) ensure the best performance and quality in Korea with the GS Certification, SW Certification for Administrative Business, New SW Product Award, and Presidential Award of Korea SW Grand Prize.
Main accomplishments
1. Strengthen Posco Family information competitiveness
- Provide useful information quickly and accurately
- Share information through division of labor and cooperation
- Use information effectively and increase business productivity using up to date information technology
2. Posco Family mental and cultural information settlement
- Establish the basis for the development of mental information through information collection and use (formation → expansion → settlement)
- Encourage changes in management through information monitoring and feedback
3. Detect strategic insight information
- Systematic accumulation of meaningful information in Korea and abroad
- Support decision making with strategic insight information
- Find new business ideas and manage risks
4. Promote POSRI information-based analysis
- Establish efficient communication between POSRI and Families and support decision making
- Enhance POSRI field research/tasking by gathering real-time domestic and overseas information