Articles tagged with Data Mining
Comprehensive exploration of living organisms, biological systems, and life processes across all scales from molecules to ecosystems. Encompasses cutting-edge research in biology, genetics, molecular biology, ecology, biochemistry, microbiology, botany, zoology, evolutionary biology, genomics, and biotechnology. Investigates cellular mechanisms, organism development, genetic inheritance, biodiversity conservation, metabolic processes, protein synthesis, DNA sequencing, CRISPR gene editing, stem cell research, and the fundamental principles governing all forms of life on Earth.
Comprehensive medical research, clinical studies, and healthcare sciences focused on disease prevention, diagnosis, and treatment. Encompasses clinical medicine, public health, pharmacology, epidemiology, medical specialties, disease mechanisms, therapeutic interventions, healthcare innovation, precision medicine, telemedicine, medical devices, drug development, clinical trials, patient care, mental health, nutrition science, health policy, and the application of medical science to improve human health, wellbeing, and quality of life across diverse populations.
Comprehensive investigation of human society, behavior, relationships, and social structures through systematic research and analysis. Encompasses psychology, sociology, anthropology, economics, political science, linguistics, education, demography, communications, and social research methodologies. Examines human cognition, social interactions, cultural phenomena, economic systems, political institutions, language and communication, educational processes, population dynamics, and the complex social, cultural, economic, and political forces shaping human societies, communities, and civilizations throughout history and across the contemporary world.
Fundamental study of the non-living natural world, matter, energy, and physical phenomena governing the universe. Encompasses physics, chemistry, earth sciences, atmospheric sciences, oceanography, materials science, and the investigation of physical laws, chemical reactions, geological processes, climate systems, and planetary dynamics. Explores everything from subatomic particles and quantum mechanics to planetary systems and cosmic phenomena, including energy transformations, molecular interactions, elemental properties, weather patterns, tectonic activity, and the fundamental forces and principles underlying the physical nature of reality.
Practical application of scientific knowledge and engineering principles to solve real-world problems and develop innovative technologies. Encompasses all engineering disciplines, technology development, computer science, artificial intelligence, environmental sciences, agriculture, materials applications, energy systems, and industrial innovation. Bridges theoretical research with tangible solutions for infrastructure, manufacturing, computing, communications, transportation, construction, sustainable development, and emerging technologies that advance human capabilities, improve quality of life, and address societal challenges through scientific innovation and technological progress.
Study of the practice, culture, infrastructure, and social dimensions of science itself. Addresses how science is conducted, organized, communicated, and integrated into society. Encompasses research funding mechanisms, scientific publishing systems, peer review processes, academic ethics, science policy, research institutions, scientific collaboration networks, science education, career development, research programs, scientific methods, science communication, and the sociology of scientific discovery. Examines the human, institutional, and cultural aspects of scientific enterprise, knowledge production, and the translation of research into societal benefit.
Comprehensive study of the universe beyond Earth, encompassing celestial objects, cosmic phenomena, and space exploration. Includes astronomy, astrophysics, planetary science, cosmology, space physics, astrobiology, and space technology. Investigates stars, galaxies, planets, moons, asteroids, comets, black holes, nebulae, exoplanets, dark matter, dark energy, cosmic microwave background, stellar evolution, planetary formation, space weather, solar system dynamics, the search for extraterrestrial life, and humanity's efforts to explore, understand, and unlock the mysteries of the cosmos through observation, theory, and space missions.
Comprehensive examination of tools, techniques, methodologies, and approaches used across scientific disciplines to conduct research, collect data, and analyze results. Encompasses experimental procedures, analytical methods, measurement techniques, instrumentation, imaging technologies, spectroscopic methods, laboratory protocols, observational studies, statistical analysis, computational methods, data visualization, quality control, and methodological innovations. Addresses the practical techniques and theoretical frameworks enabling scientists to investigate phenomena, test hypotheses, gather evidence, ensure reproducibility, and generate reliable knowledge through systematic, rigorous investigation across all areas of scientific inquiry.
Study of abstract structures, patterns, quantities, relationships, and logical reasoning through pure and applied mathematical disciplines. Encompasses algebra, calculus, geometry, topology, number theory, analysis, discrete mathematics, mathematical logic, set theory, probability, statistics, and computational mathematics. Investigates mathematical structures, theorems, proofs, algorithms, functions, equations, and the rigorous logical frameworks underlying quantitative reasoning. Provides the foundational language and tools for all scientific fields, enabling precise description of natural phenomena, modeling of complex systems, and the development of technologies across physics, engineering, computer science, economics, and all quantitative sciences.
The American Heart Association has awarded $2 million in grants to researchers, clinicians and computer engineers to develop precision medicine projects. The grants will support the use of cloud computing and data analytics to improve cardiovascular disease prevention and treatment.
The project aims to improve the usage of digital technologies for research by developing new methods for analyzing historical newspapers. The team will use deep learning to correct errors in text recognition and enhance entity recognition using external data repositories.
The article reviews current chemical search engines, named entity recognition and text mining systems to efficiently access biomedical data. Researchers emphasize the need for structured databases to process unstructured data in scientific literature, clinical reports, and patents.
The LimTox tool provides information on drug hepatotoxicity extracted from biomedical archives, enabling efficient associations to adverse reactions. It offers keyword searches and entity-specific queries for researchers and clinicians, promoting targeted search queries and biological knowledgebase construction.
Researchers at Rice University have developed a new indoor location detection system that uses existing mobile device sensors to improve accuracy and energy efficiency. By leveraging machine learning algorithms and incorporating information from standard human movements, the system can estimate accurate locations even with noisy sensors.
A new study from Columbia Business School sheds light on the secret sauce to developing creative ideas, finding that a balance between novelty and familiarity makes an idea more creative. The researchers developed a tool that analyzes word combinations in real-time and recommends words to improve ideas.
Researchers at KAUST developed ScaleMine, a system that accelerates frequent subgraph mining (FSM) by up to ten-fold, allowing for faster analysis of large graph data. The new approach uses a two-step process to divide and conquer the search space, resulting in significant performance improvements.
Researchers from University of Extremadura develop non-destructive method using Magnetic Resonance Imaging to quantify salt content of Iberian ham and classify it according to salt penetration. This methodology enables real-time results and can be applied to other quality parameters.
Amy McGovern uses high-resolution simulations, data mining, and visualization techniques to identify precursors of tornadoes and improve warning lead times. Her goal is to reduce false alarms and increase prediction accuracy.
A novel tensor mining tool enables automated modeling in big data applications, facilitating the analysis of complex multiaspect data. This innovation addresses the challenge of extracting knowledge from massive amounts of data represented as tensors.
The FRAUDAR algorithm identifies over 4,000 suspicious Twitter accounts, including those using follower-buying services. The method sees through camouflage to reveal legitimate users from fraudulent ones.
A recent study by INFORMS found that tweets are more effective in predicting future TV ratings than Google Trends. The researchers analyzed over 1.8 billion tweets and 113 million Google searches for TV shows during the 2008-2012 seasons.
A two-and-a-half-year study of UK households found that consumers who start buying groceries online at a particular chain tend to spend more at that chain than elsewhere. Grocery retailers can attract customers to their online channel by offering personalized features and customized promotions.
A study published in Scientific Reports found that vitamin D lowers the risk of developing hyperglycaemia when taken with atypical antipsychotics like quetiapine. The researchers also confirmed this finding using mouse tests, revealing that vitamin D defends against the insulin-lowering effects of quetiapine.
The Data Miners team from INRS took second place in the Integra Gold Rush Challenge, a competition that analyzed six terabytes of vintage data to locate the next prospective gold deposit on the Integra property. The team was composed entirely of students and demonstrated the excellence of the training in economic geology offered at INRS.
A new study proposes using data mining tools to identify patterns in QS data that can inform users' decisions on diet and exercise. This approach has the potential to reveal ways to improve personal well-being without compromising data privacy.
A new research method has been developed by Virginia Tech computer scientists to identify and prevent stealthy attacks on complex computer systems. The approach uses matrix-based pattern recognition to analyze execution paths and detect anomalies with low false positive rates.
Researchers from the University of Kent have developed an automated data mining system that can mimic human expert classification of potentially illegal elephant ivory with high accuracy. The system has shown a 93% accuracy rate and is expected to significantly increase the detection of illegal ivory sales on eBay.
Researchers created a publicly available website, neuroelectro.org, to collect and standardize data on neuronal function. The site enables the comparison of physiological information across different types of neurons, promoting new methods of analysis.
Academic researchers are mining social media data to learn about online and offline human behavior, but flaws in studies point to need for more aware analysis methods. The study highlights issues such as user demographics, data filtering, platform design, spam bots, and biased results.
The 2014 Semantic Web Challenge was won by three teams: Mining the Web of Linked Data with RapidMiner, Enabling Live Exploration on The Graph of Things, and DIVE into the Event-Based Browsing of Linked Historical Media. The Big Data Track was won by Extending Tables with Data from over a million websites.
A team of UT Arlington Computer Science and Engineering students won the NTx Apps Challenge with a smart traffic light network that analyzes traffic conditions in real time. The GridLock system enhances traffic flow, reducing congestion and jams.
Researchers at Cornell University have developed a new method called 'data smashing' that enables automated discovery without human intervention, opening doors to complex observations and expert-driven analysis.
A five-year project will create a large, distributed infrastructure called LearnSphere to securely store data on student learning. By accessing over 550 datasets, course developers and instructors can improve teaching and learning through data-driven design.
A new tool called KnIT has shown promise in mining all public medical literature and generating hypotheses that can lead to breakthroughs in cancer research. The tool was tested on the protein p53, which is a critical tumor suppressor protein, and accurately predicted the existence of proteins that modify it.
Researchers analyzed five million images associated with 48 brands to identify visual concepts and keywords. The study complements online text data analysis and suggests new directions for computer vision in electronic commerce.
A team of researchers at Rensselaer Polytechnic Institute is exploring a novel two-stage approach to harnessing petabyte data, combining cloud computing with precise computational systems. They aim to develop algorithms and methods for extracting knowledge from massive amounts of data.
Researchers used clustering techniques to analyze 300 patient cases and identify specific risk factors associated with heart attack risk. Key findings include the importance of age, gender, and lifestyle habits in determining cardiac risk levels.
A new computational technique allows predicting infectious avian influenza strains based on protein sequences and physicochemical characteristics. The 'A2H' system has successfully validated its predictions against known strains of bird flu and those that are infectious to people.
Lian Duan's new algorithm cluster improves adverse drug effect detection by 23.83% in simulated medical outcomes dataset. NJIT has a history of assisting New Jersey physicians with electronic health information technology, and Duan's research has applications in various industries.
Researchers are mining data from pediatric intensive care units to find regularities and patterns that can aid doctors in diagnosing and predicting medical episodes. The team plans to incorporate these findings into real-time sensors to see if they help doctors.
A new study by Columbia Business School researchers uses text mining to analyze consumer-generated content, providing valuable insights on market structure and competitive landscape. The method can be used to monitor market positions over time and assess the effectiveness of marketing campaigns.
Researchers developed visual data mining software to detect subtle city-specific features in Google Street View images. The software identified unique elements such as cast-iron balconies in Paris and fire escapes in New York City, which can be used for computational geography tasks.
Fourteen international teams win $4.8 million in grants to investigate 'big data' techniques for humanities and social sciences research. The projects cover topics such as music, Egyptian mummies, and human rights abuses.
Virginia Tech's Virginia Tech Knowledge Networks (VTKN) accelerates science, technology education & research by leveraging data mining & authorship network visualizations. The platform provides community memory for faculty members to build on prior work, revealing potential collaborations and novel outcomes.
The San Diego Supercomputer Center is launching a new data-intensive supercomputer system named Gordon, which will have 250 trillion bytes of flash memory and 64 I/O nodes. This new system aims to solve the challenge of storing valuable data and accelerating scientific discovery by providing faster speeds and massive storage capabilities.
Researchers used data mining to analyze the corrosion-resistant properties of Alloy 22, a key material for nuclear waste containment. They found that the alloy can predict future corrosion patterns under similar environmental conditions.
Researchers at Northwestern University have found that using electronic medical records to identify patients with diseases can be faster and cheaper than recruiting thousands of participants. The study used data from five national sites to accurately identify patients with five types of diseases, achieving accuracy rates of 73-98%.
Researchers at Concordia University have developed a novel technique for identifying anonymous email authors with high accuracy. By analyzing unique patterns in a suspect's emails, they can create a 'write-print' that is like a fingerprint, allowing investigators to determine the author's gender, nationality, and education level.
Researchers develop system to analyze patient data, therapist interactions, for better understanding of depression causes and prevention strategies. The approach balances individual differences with observed similarities in behavior and response to treatment.
A Virginia Tech expert will assess all aspects of visitor experiences in state parks to provide actionable recommendations. The study aims to improve park operations and services through data-driven decision making.
Mohammed J. Zaki, a Rensselaer Polytechnic Institute professor, has been selected for the 2010 HP Labs Innovation Research Program for his groundbreaking work on graph patterns and link analysis. His research aims to uncover hidden connections between entities and data, enabling more comprehensive insights into complex phenomena.
A Stanford University study finds that children of gay and married couples have lower grade-repetition rates than their peers in other family environments. This research challenges common assertions that children of same-sex couples cannot thrive, providing new data to the debate on gay marriage.
A University of Utah computer scientist has devised a new method to simplify and speed up data mining, allowing for the analysis of high-dimensional data. The new approach can handle larger datasets than previous methods, making it useful for various applications in natural and social sciences.
The study compared two chemical name dictionaries and found that automatic curation with Chemlist outperformed manual curation with ChemSpider. The Chemlist dictionary achieved a higher recall and better F-score, while ChemSpider's precision was higher after filtering and disambiguation.
Research shows consumers develop specific rules for sharing personal data online, influencing retailers' data mining practices. Understanding these rules is crucial to improving online experiences and products.
Researchers developed a system to extract subjective expressions from web pages, scoring them for subjectivity and indicating whether they express personal opinions or marketing speak. The method outperformed general search engines in detecting personal opinion pages across four categories.
Researchers mapped brain activity associated with speech sounds and voices to identify unique 'neural fingerprints' in listeners. This breakthrough could improve computer systems for automatic speech recognition, revealing a less hierarchical processing of speech across the brain.
Researchers Edgar de Graaf and Jeroen De Knijf analyzed patterns in web surfing behavior to reduce the number of results and improve efficiency in data mining. They developed methods to detect relevant patterns quickly and effectively within large quantities of semi-structured data.
A study finds that many girls with exceptional math talent exist but are rarely identified due to a lack of respect for math and role models. The US pipeline for nurturing top math talent is badly broken, with 80% of female faculty hired at top research universities born in other countries.
Researchers used data mining to identify subtle differences in movement between rats with ALS mutations and controls, predicting the onset of disease two months early. This novel approach may enable testing of therapies to delay or prevent disease.
Researchers applied data mining to a human resources database to discover patterns that can improve business efficiency and profits. By analyzing skills, qualifications, employment history, and interactions between personnel, businesses can predict natural staff turnover, morale changes, and employee performance.
University of Houston and Lunar Planetary Institute researchers use data mining and machine learning to analyze Martian surface patterns. They develop new computational tools to characterize large portions of the Martian landscape, aiding understanding of past water activity on Mars.
Researchers at UCI used text mining to analyze 330,000 NYC Times articles, identifying topics and trends. This technology can be applied to various fields, such as medical research and law.
The new tool uses natural language processing and historical knowledge to extract entities from historical documents, allowing users to quickly locate specific individuals or locations. The system will enable users to browse lists of entities and their frequencies within individual documents and the collection as a whole.
The workshop explored the government's use of commercial data for homeland security, raising important privacy concerns. Stevens' Wright discussed her work on PORTIA project, which addresses handling of sensitive information in a networked world and provides cryptographically strong methods for maintaining data privately.
The National Centre for Text Mining provides a groundbreaking service for the academic community, utilizing natural language processing and data mining to uncover hidden patterns and associations. This innovative approach has already shown promise in fields like drug discovery and predictive toxicology.
The project aims to evaluate the effectiveness of data-mining tools in identifying threat levels and national security interests. The UC Riverside team, including graduate students and staff, will work closely with experts from Bell Laboratories and Lucent Technologies.
Researchers have created a computational tool to mine genomic data and identify biologically meaningful gene regulatory networks. The tool uses a probabilistic framework that integrates data from various sources, including microarrays, DNA sequences, and protein-protein interactions.
The researchers report that the rough sets method offers significant advantages in intrusion detection, including the ability to work with missing values and imprecise data. The team's study found an average classification accuracy rate of 75.68% for rough sets, compared to 69.78% for neural nets and 51.16% for inductive learning.