Agnieszka Dardzinska

Association Action Rules

Action rules describe possible transitions of objects from one state to another with respect to a distinguished attribute. Previous research on action rule discovery usually required the extraction of classification rules before constructing any action rule. This paper gives anew approach for generating association-type action rules. The notion of frequent action sets and Apriori-like strategy generating them is proposed. We introduce the notion of a representative action rules and give an algorithm to construct them directly from frequent action sets. Finally, we introduce the notion of a simple association action rule, the cost of association action rule, and give a strategy to construct simple association action rules of a lowest cost.

Ontology-based distributed autonomous knowledge systems

Traditional query processing usually requires that users fully understand the database structure and content to issue a query. Due to the complexity of the database applications and the variety of user needs, the so-called global queries are introduced which traditional query answering systems cannot handle. Query posed to a database D is global if minimum one of its attributes is missing in D while it occurs in other databases. Definitions of a missing attribute in D can be extracted from other databases and shared with D. To handle semantics inconsistencies between the same attributes used at different sites, task ontologies are used as a communication bridge between them. These inconsistencies can be caused either by different granularity levels or by different interpretations of the same attribute. As the final outcome of this research, a rough query answering system based on distributed data mining is presented.

Cursive Letters Language Processing: Muqla Model and Toeplitz Matrices Approach

This article presents a new model (based on Muqla idea) used in recognition of cursive letters. It is used for creating feature vectors based on Toeplitz matrices. As the approach is used for the first time, no comparison had been made from modelling point of view. However, the way of feature vectors extraction in this approach showed its simplicity and may find its practical applications in processing and verifying both written and spoken texts. The suggested algorithm makes it possible to reduce the limiting factors [1] in recognition as it is easy to extend it to cover more vocabulary, syntax or semantic information.

HANDLING SEMANTIC INCONSISTENCIES IN QUERY ANSWERING BASED ON DISTRIBUTED KNOWLEDGE MINING

Traditional query processing provides exact answers to queries. It usually requires that users fully understand the database structure and content to issue a query. Due to the complexity of the database applications, the so-called global queries can be posed which traditional query answering systems cannot handle. Query posed to a database D is global if a minimum of its attributes is missing in D while it occurs in other databases. In this paper a query answering system based on distributed data mining is presented to rectify these problems. Task ontologies are used as a tool to handle semantic inconsistencies between sites.

Handling Semantic Inconsistencies in Distributed Knowledge Systems Using Ontologies

Traditional query processing provides exact answers to queries. It usually requires that users fully understand the database structure and content to issue a query. Due to the complexity of the database applications, the so called global queries can be posed which traditional query answering systems can not handle. In this paper a query answering system based on distributed data mining is presented to rectify these problems. Task ontologies are used as a tool to handle semantic inconsistencies between sites.

Natural language processing: word recognition without segmentation

In an earlier article about the methods of recognition of machine and hand-written cursive letters, we presented a model showing the possibility of processing, classifying, and hence recognizing such scripts as images. The practical results we obtained encouraged us to extend the theory to an algorithm for word recognition. In this article, we introduce our ideas, describe our achievements, and present our results of testing words for recognition without segmentation. This would lead to the possibility of applying the methods used in this work, together with other previously developed algorithms to process whole sentences and, hence, written and spoken texts with the goal of automatic recognition.

Comparative evaluation of the different data mining techniques used for the medical database

Abstract Data mining is the upcoming research area to solve various problems. Classification and finding association are two main steps in the field of data mining. In this paper, we use three classification algorithms: J48 (an open source Java implementation of C4.5 algorithm), Multilayer Perceptron - MLP (a modification of the standard linear perceptron) and Naïve Bayes (based on Bayes rule and a set of conditional independence assumptions) of the Weka interface. These classifiers have been used to choose the best algorithm based on the conditions of the voice disorders database. To find association rules over transactional medical database first we use apriori algorithm for frequent item set mining. These two initial steps of analysis will help to create the medical knowledgebase. The ultimate goal is to build a model, which can improve the way to read and interpret the existing data in medical database and future data as well.

Query Answering System as a Tool in Incomplete Distributed Information System Optimization Process

We assume there is a group of connected distributed information systems (DIS). They work under the same ontology. Each information system create its own knowledgebase. Values of attributes in information system \( S \) form atomic expressions of a language used for communication with others. Collaboration among systems is initiated when one of them (called a client) is asked to resolve a query containing nonlocal attributes for \( S \). In such case, the client has to ask for help other information systems to have that query answered. As the result of its request, knowledge is extracted locally in each information system and sent back to the client. The outcome of this step is a knowledgebase created at the client site, which can be used to answer given query. In this paper we present a method of identifying which information system is semantically the closest to client.

Queries for Detailed Information System Selection

In this paper we assume there is a group of connected information systems forming distributed information system (DS). They work under the same ontology. At the same time, each information system has its own knowledge base. Values of attributes in each information system S form atomic expressions of a language used for communication with others. Collaboration among them is initiated when one of information system S is asked by user to resolve a query containing some nonlocal attributes for S. Therefore it has to contact other information systems to obtain additional, helpful knowledge for finding finally objects satisfying given query. Because there is a set of different information systems connected with a given one, we have to decide which of them is the closest with its knowledge, and which one should be selected by user, for further investigation.

Comprehensive Review of Classification Algorithms for Medical Information System

Nowadays, the Internet and information systems become an integral part of everyday life. The trend of using advanced recommendation systems is still growing in various areas, also in medicine. Two of the diseases where diagnosis is a big problem for specialists are colon disease and Crohn’s disease. The course of the disease strongly resembles other diseases in the large intestine, so it became extremely important to help doctors and find symptoms that would clearly indicate the colon disease, excluding others. In order to find rules that distinguish these two diseases, together data mining and statistical methods were mixed and used.