Mining Target-Oriented Sequential Patterns with Time-Intervals

📝 Abstract

A target-oriented sequential pattern is a sequential pattern with a concerned itemset in the end of pattern. A time-interval sequential pattern is a sequential pattern with time-intervals between every pair of successive itemsets. In this paper we present an algorithm to discover target-oriented sequential pattern with time-intervals. To this end, the original sequences are reversed so that the last itemsets can be arranged in front of the sequences. The contrasts between reversed sequences and the concerned itemset are then used to exclude the irrelevant sequences. Clustering analysis is used with typical sequential pattern mining algorithm to extract the sequential patterns with time-intervals between successive itemsets. Finally, the discovered time-interval sequential patterns are reversed again to the original order for searching the target patterns.

💡 Analysis

A target-oriented sequential pattern is a sequential pattern with a concerned itemset in the end of pattern. A time-interval sequential pattern is a sequential pattern with time-intervals between every pair of successive itemsets. In this paper we present an algorithm to discover target-oriented sequential pattern with time-intervals. To this end, the original sequences are reversed so that the last itemsets can be arranged in front of the sequences. The contrasts between reversed sequences and the concerned itemset are then used to exclude the irrelevant sequences. Clustering analysis is used with typical sequential pattern mining algorithm to extract the sequential patterns with time-intervals between successive itemsets. Finally, the discovered time-interval sequential patterns are reversed again to the original order for searching the target patterns.

📄 Content

International journal of computer science & information Technology (IJCSIT) Vol.2, No.4, August 2010

DOI : 10.5121/ijcsit.2010.2410 113 MINING TARGET-ORIENTED SEQUENTIAL PATTERNS WITH TIME-INTERVALS Hao-En Chueh Department of Information Management, Yuanpei University, Hsinchu City, Taiwan hechueh@mail.ypu.edu.tw ABSTRACT A target-oriented sequential pattern is a sequential pattern with a concerned itemset in the end of pattern. A time-interval sequential pattern is a sequential pattern with time-intervals between every pair of successive itemsets. In this paper we present an algorithm to discover target-oriented sequential pattern with time-intervals. To this end, the original sequences are reversed so that the last itemsets can be arranged in front of the sequences. The contrasts between reversed sequences and the concerned itemset are then used to exclude the irrelevant sequences. Clustering analysis is used with typical sequential pattern mining algorithm to extract the sequential patterns with time-intervals between successive itemsets. Finally, the discovered time-interval sequential patterns are reversed again to the original order for searching the target patterns. KEYWORDS Data Mining, Target-Oriented Sequential Pattern, Time-Interval, Clustering Analysis

1. INTRODUCTION Data mining (sometimes called knowledge discovery in databases) is the process of finding correlations or patterns among dozens of fields in large relational databases [4, 9, 12]. The primary tasks of data mining include association analysis, clustering analysis, classification, pattern recognition, prediction, etc. It has been widely used in business and engineering fields to discover useful information but previously unknown [15, 16, 20, 23]. Discovering sequential patterns is one of the most important tasks in data mining. The procedure of data mining is focused on mining sequential patterns is the task of finding frequently occurring patterns related to time or other sequences from a sequence database [1]. An example of a sequential pattern is “A customer who bought a digital camera will buy an extra memory card later”. This kind of data mining task is very useful in retail business to assist decision makers in making marketing strategies [4, 9, 12, 15, 23]. Up to now, many mining sequential patterns algorithms have been proposed [1, 3, 5, 10, 14, 15, 18, 21], and most algorithms only focus on the order of the itemsets, but ignore the time- intervals between itemsets. However, in retail business, a sequential pattern with time-intervals between itemsets is more valuable than a traditional sequential pattern without any time information. An example of a sequential pattern with time-interval between itemsets is “A customer who bought a digital camera will buy an extra memory card within one week”. Clearly, the time-intervals between itemsets can offer useful information for retail business to provide the correct products to their customers at the right time. Therefore, some researches start to focus on discovering the sequential patterns with time-intervals between itemsets [2, 6, 8, 11, 13, 17, 18, 19, 23]. This kind of sequential pattern is called as time-interval sequential pattern. International journal of computer science & information Technology (IJCSIT) Vol.2, No.4, August 2010 114

To obtain time-intervals between every pair of successive itemsets, clustering analysis is used in this paper to automatically generate the suitable time partitions between frequent occurring pairs of successive itemsets, and then uses these obtained time-intervals to extend typical algorithms to discover the time-interval sequential patterns [8]. In addition, for most marketing decision makers, they usually need to know the happening order of some concerned itemsets. This kind of sequential pattern ca ne called as target-oriented sequential pattern [7]. To this end, the original sequences are reversed so that the last itemsets can be arranged in front of sequences. The contrasts between the reversed sequences and the concerned itemset are then used to exclude the irrelevant sequences. These reversed sequences are analyzed to extract sequential patterns with time-interval between successive itemsets. Finally, the discovered time-interval sequential patterns are reversed again to the original order for searching the target patterns. An algorithm to discover target-oriented sequential pattern with time-intervals between successive itemsets is presented in this paper. The rest of this paper is organized as follows: Some researches related to target-oriented sequential patterns and time-interval sequential patterns are reviewed in section 2. The algorithm to mine target-oriented sequential patterns with time-intervals between successive itemsets is presented in section 3. A s

This content is AI-processed based on ArXiv data.