SPEED : Mining Maximal Sequential Patterns over Data Streams
#1

presented by:
Chedy Ra¨ıssi,Pascal Poncelet,Maguelonne Teisseire

ABSTRACT
Many recent real-world applications, such as network traf-_c monitoring, intrusion detection systems, sensor networkdata analysis, click stream mining and dynamic tracing of_nancial transactions, call for studying a new kind of data.Called stream data, this model is, in fact, a continuous, potentiallyin_nite ow of information as opposed to _nite,statically stored data sets extensively studied by researchersof the data mining community. An important applicationis to mine data streams for interesting patterns or anomaliesas they happen. For data stream applications, the volumeof data is usually too huge to be stored on permanentdevices, main memory or to be scanned thoroughly morethan once. We thus need to introduce approximations whenexecuting queries and performing mining tasks over rapiddata streams. In this paper we propose a new approach,called Speed (Sequential Patterns E_cient Extraction inData streams), to identify frequent maximal sequential patternsin a data stream. To the best of our knowledge this isthe _rst approach de_ned for mining sequential patterns instreaming data. The main originality of our mining methodis that we use a novel data structure to maintain frequentsequential patterns coupled with a fast pruning strategy. Atany time, users can issue requests for frequent maximal sequencesover an arbitrary time interval. Furthermore, ourapproach produces an approximate support answer with anassurance that it will not bypass a user-de_ned frequencyerror threshold. Finally the proposed method is analyzedby a series of experiments on di_erent datasets
INTRODUCTION
Recently, the data mining community has focused on a newchallenging model where data arrives sequentially in theform of continuous rapid streams. It is often referred to asdata streams or streaming data. Since data streams are continuous,high-speed and unbounded, it is impossible to mineassociation rules by using algorithms that require multiplescans. As a consequence new approaches were proposed tomaintain itemsets [7, 4, 3, 6, 13]. Nevertheless, according tothe de_nition of itemsets, they consider that there is no limitationon items order. In this paper we consider that itemsare really ordered into the streams, therefore we are interestedin mining sequences rather than itemsets. We proposea new approach, called Speed (Sequential Patterns E_cientExtraction in Data streams), to mine sequential patterns ina data stream. The main originality of our approach is thatwe use a novel data structure to incrementally maintain frequentsequential patterns (with the help of tilted-time windows)coupled with a fast pruning strategy. At any time,users can issue requests for frequent sequences over an arbitrarytime interval. Furthermore, our approach produces anapproximate support answer with an assurance that it willnot bypass a user-de_ned frequency thresholds.The remainder of the paper is organized as follows. Section2 goes deeper into presenting the problem statement. InSection 3 we propose a brief overview of related work.TheSpeed approach is presented in Section 4. Section 5 reportsthe result of our experiments. In Section 6, we conclude thepaper.

download full report
http://hal.archives-ouvertes.fr/docs/00/...-SPEED.pdf
Reply

Important Note..!

If you are not satisfied with above reply ,..Please

ASK HERE

So that we will collect data for you and will made reply to the request....OR try below "QUICK REPLY" box to add a reply to this page
Popular Searches: sequential diagram of metical shop menegement, sequential direct online starter, thesis on pbfmcsp prefix based fast mining of closed sequential patterns, mining data streams, sequential starter circuits, advanced plastic canvas patterns, speed detection camera system using image processing techniques on video streams,

[-]
Quick Reply
Message
Type your reply to this message here.

Image Verification
Please enter the text contained within the image into the text box below it. This process is used to prevent automated spam bots.
Image Verification
(case insensitive)

Possibly Related Threads...
Thread Author Replies Views Last Post
  Block Chain and Data Science jntuworldforum 0 7,955 06-10-2018, 12:15 PM
Last Post: jntuworldforum
  Data Encryption Standard (DES) seminar class 2 9,331 20-02-2016, 01:59 PM
Last Post: seminar report asees
  Skin Tone based Secret Data hiding in Images seminar class 9 6,980 23-12-2015, 04:18 PM
Last Post: HelloGFS
Brick XML Data Compression computer science crazy 2 2,377 07-10-2014, 09:26 PM
Last Post: seminar report asees
  Data Security in Local Network using Distributed Firewalls computer science crazy 10 14,783 30-03-2014, 04:40 AM
Last Post: Guest
  GREEN CLOUD -A Data Center Approach computer topic 0 1,530 25-03-2014, 10:13 PM
Last Post: computer topic
  3D-OPTICAL DATA STORAGE TECHNOLOGY computer science crazy 3 8,499 12-09-2013, 08:28 PM
Last Post: Guest
  Security in Data Warehousing seminar surveyer 3 9,833 12-08-2013, 10:24 AM
Last Post: computer topic
  data warehousing concepts project topics 7 7,112 05-02-2013, 12:00 PM
Last Post: seminar details
Star DATA MINING AND WAREHOUSE seminar projects crazy 2 3,357 05-02-2013, 12:00 PM
Last Post: seminar details

Forum Jump: