کاوش کارامد قوانین روابط فازی از جریان های فراگیر داده ها Efficient mining fuzzy association rules from ubiquitous data streams

۱٫ Introduction Recent emerging applications, such as network traffic monitoring, sensor network data analysis, web click stream mining, power consumption measurement, and dynamic tracing of stock market fluctuations, call for studying a new kind of data. This is called stream data, which can be continuous, potentially infinite flow of information, as opposed to finite, statically stored data sets. Stream Data Mining is the process of extracting knowledge structures from continuous, and rapid data records. The dissemination of data streams systems, wireless networks and mobile/handheld devices motivates the need for an efficient data analysis tool capable of gaining insights about these continuous data streams [1]. Ubiquitous data streams mining (UDM) is the process of pattern discovery on mobile, embedded and ubiquitous devices. It represents the next generation of data mining systems, that will support the intelligent and time-critical information needs of mobile users, and will facilitate ‘‘anytime, anywhere’’ data mining. Fuzzy logic is a type of logic used in artificial intelligence. It is referred to as a multi-valued logic.Instead of having two values (true and false), there are a continuum of possible truth values [2]. In fuzzy logic, every proposition is a statement that is assigned a number between 0 (false) and 1 (true), such a statement is called a fuzzy proposition. Fuzzy logic provides a powerful tool to categorize a concept in an abstract way by introducing vagueness. Many data streams applications exist, that require association rule mining, such as network traffic monitoring and web click streams analysis. These applications’ goal is to discover important associations among items as the presence of some items will imply the presence of others. Fuzzy association rule approach could combine data mining results with human expertise and background knowledge, in the form of rules, to attain labeled classes for classification of data streams. Another advantage of the fuzzy logic approach is that it gives classification results, which include a degree of probability. This paper demonstrates the effectiveness of Fuzzy Association Rules Mining from Ubiquitous Data Streams. This will be revealed in the coming sections. For this purpose, the remaining part of the paper is organized as follows: the work related to Fuzzy Association rules mining and ubiquitous data streams mining are reviewed and summarized in Section 2. An efficient fuzzy association rules mining technique from ubiquitous data streams is proposed in Section 3, and its complexity is analyzed in Section 4. Moreover,experimental results are discussed in Section 5. The paper is concluded and future research issues are presented, in Section 6. ۲٫ Related work This paper belongs to different inter-related research fields. The two main related topics of this work are presented: Ubiquitous Data Streams Mining Techniques that can effectively analyze continuously streaming data and Fuzzy association rules mining algorithms. These two fields will be surveyed. ۲٫۱٫ Ubiquitous data streams mining The approach, based on finite statically stored data sets, is not satisfactory in several applications. These include wireless network analysis, intrusion detection, stock market analysis, sensor network data analysis, and, in general, any setting in which every information available should be used to make an immediate decision. Such situations demand new algorithms, that are able to cope with evolutions of data as shown in Table 1 [3].