The problem of mining frequent set is a key issue in data mining. In this paper, a new method of miningfrequent set based on the frequent link is proposed. The algorithm constructs alternate frequent link from the tra...The problem of mining frequent set is a key issue in data mining. In this paper, a new method of miningfrequent set based on the frequent link is proposed. The algorithm constructs alternate frequent link from the transac-tion, the alternate link is yielded by adding up the alternate frequent link which constructed by scanning the transac-tion database in proper order. The frequent link that comprises all the information is constructed with the frequentnode which is selected according requirement. Our algorithm need to scan the transaction database only once and easysupervises the change of frequent set in order to guarantee the right of association rule.展开更多
Packet analysis is very important in our digital life. But what protocol analyzers can do is limited because they can only process data in determined format. This paper puts forward a solution to decode raw data in an...Packet analysis is very important in our digital life. But what protocol analyzers can do is limited because they can only process data in determined format. This paper puts forward a solution to decode raw data in an unknown format. It is certain that data can be cut into packets because there are usually characteristic bit sequences in packet headers. The key to solve the problem is how to find out those characteristic sequences. We present an efficient way of bit sequence enumeration. Both Aho-Corasick (AC) algorithm and data mining method are used to reduce the cost of the process.展开更多
文摘The problem of mining frequent set is a key issue in data mining. In this paper, a new method of miningfrequent set based on the frequent link is proposed. The algorithm constructs alternate frequent link from the transac-tion, the alternate link is yielded by adding up the alternate frequent link which constructed by scanning the transac-tion database in proper order. The frequent link that comprises all the information is constructed with the frequentnode which is selected according requirement. Our algorithm need to scan the transaction database only once and easysupervises the change of frequent set in order to guarantee the right of association rule.
文摘Packet analysis is very important in our digital life. But what protocol analyzers can do is limited because they can only process data in determined format. This paper puts forward a solution to decode raw data in an unknown format. It is certain that data can be cut into packets because there are usually characteristic bit sequences in packet headers. The key to solve the problem is how to find out those characteristic sequences. We present an efficient way of bit sequence enumeration. Both Aho-Corasick (AC) algorithm and data mining method are used to reduce the cost of the process.
