not enough. When you understand what you see, seeing becomes believing. A while ago scientists A reliable visual data mining system must provide estimated error or accuracy of the projected informa- tion for each step of the mining process. This error infor- discovered that seeing and understanding together mation can compensate for the deficiency that imprecise enable humans to glean knowledge and deeper insight analysis of data visualization can cause. from large amounts of data. The approach integrates the A reusable visual data mining system must be adapt- human minds exploration abilities with the enormous able to a variety of systems and environments to reduce processing power of computers to form a powerful the customization effort, provide assured performance, knowledge discovery environment that capitalizes on and improve system portability. the best of both worlds. The technology builds on visual A practical visual data mining system must be gener- and analytical processes developed in various disciplines ally and widely available. The quest for new knowledge including scientific visualization, data mining, statistics, or deeper insights of existing knowledge cannot be and machine learning with custom extensions that han- planned. It may mean a portable system through dle very large, multidimensional, multivariate data sets. telelinks or an embedded (local) system within the The methodology is based on both functionality that information domain. This requires that the knowledge characterizes structures and displays data and human received from one domain adapt to another domain capabilities that perceive patterns, exceptions, trends, through physical means or electronic connections. and relationships. Here Ill define the vision, present the Finally, a complete visual data mining system must state of the art, and discuss the future of a young disci- include security measures to protect the data, the newly pline called visual data mining. discovered knowledge, and the users identity because of various social issues. The vision So far Ive ignored discussing the underlying visual- The vision of a visual data mining system stems from ization and mathematical techniques of visual data min- the following principles: simplicity, user autonomy, reli- ing. This is partly because of the space limit and partly ability, reusability, availability, and security. because of the steady and incremental technological A visual data mining system must be syntactically sim- advancements in the field of visual data mining. You can ple to be useful. Simple doesnt mean trivial or nonpow- find samples of the latest technologies in this special erful. Simple to learn means use of intuitive and friendly issue. Although no one involved in this exciting field has input mechanisms as well as instinctive and easy-to-inter- all the technical solutions today, everyone is fully aware pret output knowledge. Simple to apply means an effec- of the grand challenges ahead. tive discourse between humans and information. Simple to retrieve or recall means a customized data structure to Current state of the art facilitate fast and reliable searches. Simple to execute Visualization has been used routinely in data mining means a minimum number of steps needed to achieve as a presentation tool to generate initial views, navigate the results. In short, simple means the smallest, func- data with complicated structures, and convey the tionally sufficient system possible. results of an analysis. Generally, the analytical meth- A genuine visual data mining system must not impose ods themselves dont involve visualization. The loose- knowledge on its users, but instead guide them through ly coupled relationships between visualization and the mining process to draw conclusions. Humans should analytical data mining techniques represent the major- study the visual abstractions and gain insight instead of ity of todays state of the art in visual data mining. The accepting an automated decision. process sandwich strategy, which interlaces analytical
processes with graphic visualization, penalizes both time. This provides continuous interactions between procedures with each others deficiencies and limita- man and machine during the data unfolding process. tions. For example, because an analytical process cant Because of the layered design of the visualization and analyze multimedia data, we have to give up the clustering processes, its considered a loosely coupled strengths of visualization to study movies and music in visual mining system. a visual data mining environment. Rohrer, Ebert, and Sibert describe a shape-based visu- Perhaps a stronger visual data mining strategy lies in alization system to support data mining of text. The text tightly coupling the visualizations and analytical process- information is mapped to document vectors before its es into one data mining tool. Letting human visualiza- visualized using implicit surface modeling techniques. tion participate in an analytical process decision-making The system supports querying articles of a corpus by remains a major challenge. Certain mathematical steps matching the vectors shape through visualization. within an analytical procedure may be substituted by Zoom views support data drilling of the document text. human decisions based on visualization to allow the same analytical procedure to analyze a broader scope of Conclusion information. Visualization supports humans in dealing This issue showcases an exciting field where people with decisions that can no longer be automated. This turn seeing into knowing, believing, and eventually results in a tightly coupled visual data mining environ- human insights. I believe the vision defined here can be ment that truly takes advantage of the strengths of all reached and the proposed tasks accomplished. As the worlds. articles in this issue show, both the loosely and tightly coupled visual data mining systems perform well in cer- The future tain domains and environments. Scientists and engi- All signs indicate that the field of visual data mining neers will continue to explore new ground and find new will continue to grow at an even faster pace in the future. applications in this young discipline. As for the future, In universities and research labs, visual data mining will I see the advancement of visual data mining resembling play a major role in physical and information sciences the rapid growth of personal computers in our society. in the study of even larger and more complex scientific The active participation of humans and the decisions data sets. It will also play an active role in nontechnical based on visualization combine the art of human intu- disciplines to establish knowledge domains to search for ition and the science of mathematical deduction, for- answers and truths. For example, there may exist stan- ever changing the landscape of data analysis. dard man pages for our favorite visual data mining functions on our Unix system. An advanced form of scat- Acknowledgments terplot matrix may substitute for the use of covariance The Pacific Northwest National Laboratory is operat- and regression in statistics studies. National standards ed for the US Department of Energy by Battelle Memo- will be developed to govern the functionality and rial Institute under contract DE-AC06-76RLO 1830. resources of visual data mining. In industries and households across the country, visu- al data mining will be embedded in public utilities and home appliances. Many searching referencessuch as Pak Chung Wong is a senior the yellow pages, dictionaries, and even newspapers research scientist in the Synthesis, will have visual mining capability. There may be com- Analysis, and Visualization of Infor- puter chips dedicated to support visual data mining mation group at the Pacific North- activities. The term visual data mining will be included west National Laboratory in in school textbooks and literature. Audio- or haptic- Richland, Washington, where he per- based substitutes will help the visually impaired. Our forms research and development on imagination is the only limit of the future. scientific computation and information technology. His research interests include visualization, data mining, sci- About the articles entific data abstraction, steganography, and wavelets. He This special issue on visual data mining attracted received a PhD in computer science from the University of high-quality submissions from England, Germany, and New Hampshire. the United States. All of the articles presented interest- ing and promising research in visual data mining. Unfor- Readers may contact Wong at the Pacific Northwest tunately, theres only room to include a few articles. National Laboratory, 902 Battelle Blvd., P.O. Box 999, Hinneburg, Keim, and Wawryniuk introduce a novel MSIN: K7-28, Richland, WA 99352, e-mail clustering algorithm on large amounts of high-dimen- pak.wong@pnl.gov. sional data. Visualization techniques instead of auto- mated decisions guide the recursive partitioning of the new clustering algorithm. This is a major step towards the goal of a tightly coupled visual data mining envi- ronment. Ribarsky et al. present a clustering algorithm for very large data sets. The new technique enables clustering of large amounts of data with a fast interactive response
Recommendation System With Automated Web Usage Data Mining by Using K-Nearest Neighbour (KNN) Classification and Artificial Neural Network (ANN) Algorithm
(Philosophy and Medicine 130) Peter J. Cataldo, Dan O'Brien - Palliative Care and Catholic Health Care - Two Millennia of Caring For The Whole Person-Springer International Publishing (2019) PDF