Sequentials

Sequential pattern mining is a data mining approach that is concerned with finding statistically relevant patterns between data examples where the values are delivered in a sequence. There are several key traditional computational problems addressed within this field. These include building efficient databases and indexes for sequence information, extracting the frequently occurring patterns, comparing sequences for similarity, and recovering missing sequence members. In general, sequence mining problems can be classified as string mining, which is typically based on string processing algorithms, and itemset mining, which is typically based on association rule learning. String mining has to do with understanding the sequence in a data set, identifying individual regions or structural units within each sequence, and then assigning a function to each structural unit. Itemset mining is used for discovering regularities between frequently co-occurring items in large transactions. For example, by analyzing the records of parolees, a rule can be produced that reads, “If a parolee finds a full-time job within one month of being released from prison, he or she is likely to keep his appointments with his parole officer.”

Frequent sequence mining is used to discover a set of patterns shared among objects that have between them a specific order. For instance, a retail shop may possess a transaction database that specifies which products were acquired by each customer over time. In this case, the store may use frequent sequence mining to find that 40% of its customers who bought the first volume of Lord of the Rings came back to buy the second volume a month later. This kind of information may be used to support directed advertising campaigns or recommendation systems. In criminal justice, frequent sequence mining could help determine the interval between certain types of crimes and the sequence of criminal offenses by offenders involved in burglary offenses. (The question to be asked and, hopefully, answered by the crime analyst might be, after a home break-in during which a violent crime occurred, what other crimes occurred, in what sequence, and with what time lapse between those offenses?) In effect, a huge number of possible sequential patterns are hidden in databases, and it is the job of crime analysts to mine those sequential patterns.