It is the research project which is ongoing.

• Venkata Ratnam

  I would like to publish a paper on outlier mining. Please suggest some popular journals.

  Please give some data on outlier mining.

  Recent replies ⋅ Show All (2)

  • 
    Venkata Ratnam replied

    I have prepared and submitted a paper on credit card fraud detection using outlier mining

    10 days ago
• Jagannath Nalavade

  I am interested in stream data mining classification using streaming random forest algorithm, Is anybody also doing research in this field?

  Help me, how to implement this project in java and MY SQL?

  Recent replies ⋅ Show All (1)

  • 
    Mani S Prasad replied

    Do you understand this algorithm fully so that its implementaion becomes simpler . Secondly you have a test streaming source for data mining .

    18 days ago
• Özkan Canay

  For a web portal and web log db structure - which data mining method is the best to use?

  I decided on the attributes of users and web services. Users attributes are sex, age group, user type (staff, student, alumni, visitors), etc. Web services (modules like mail service, phone guideI decided on the attributes of users and web services. Users attributes are sex, age group, user type (staff, student, alumni, visitors), etc. Web services (modules like mail service, phone guide service, meal menu service, etc.) attributes (characteristics) are usage time, pages per service, etc. And also services attributes can be evaluation with users (used that service) attributes. Such as services sex, services age group, services user type, etc. Here my questions starting: a) I can not be sure, what is the name of this operation (estimation, prediction, forecasting)? b) Which statistical method should be chosen to verify the suitability of features that would be appropriate? c) How can I estimate effects on quantitative values such as number of users in the portal, number of page navigation, and session durations when a new service added to the portal. d) Which method should I use? Association mining, classification, collaborative filtering,...?

  Recent replies ⋅ Show All (7)

  • 
    Omer Ajmal replied

    Expectation Maximization algo is also suitable . i have applied it on a research project for clustering user sessions based on data available in web access logExpectation Maximization algo is also suitable . i have applied it on a research project for clustering user sessions based on data available in web access log file. [http://www.enggjournals.com/ijcse/doc/IJCSE10-02-09-111.pdf]

    Apr 19, 2012
• Abeer Esmat

  Can anyone help me,i didnt find any data for money laundering.i hope someone help me to find data on money laundering?

  My research on data mining using money laundering

  Recent replies ⋅ Show All (9)

  • 
    Toon de Laaf replied

    Well, maybe the practice can help: Mr. Geert Delrue, Detective commissioner federal cvriminal police, fraud dept., anti money laundering unit can help you. You can find him on linkedin. When you wantWell, maybe the practice can help: Mr. Geert Delrue, Detective commissioner federal cvriminal police, fraud dept., anti money laundering unit can help you. You can find him on linkedin. When you want contact: mail me: A.de.Laaf@rechtspraak.nl

    Apr 18, 2012
• Vivek Singh Raghuvanshi

  I want some suggestions on creating an automatic knowledge mining software. Can anyone help me on this?

  How to remove co-reference problem? This is the major problem.

  Recent replies ⋅ Show All (1)

  • 
    Roel Hogervorst replied

    You will have to specify the criteria and databases first, this is a to broad question to answer

    Mar 30, 2012
• Sreeraj m.s

  Specificity and Sensitivity

  What is the meaning of specificity and sensitivity that has the value of 1. Is it a good condition? Will it help for a good system?

  Recent replies ⋅ Show All (1)

  • 
    Ian Clements replied

    Specificity is when the test responds to the particular item rather than any other with which it may be confused with. Sensitivity is when it picks up any occurrence and doesn't when there's none. SoSpecificity is when the test responds to the particular item rather than any other with which it may be confused with. Sensitivity is when it picks up any occurrence and doesn't when there's none. So a '1' for both is perfect: the item is there and no other; that there always is that item there rather than non (false positive). Most are somewhat below this.

    Apr 11, 2012
• Heidi Imker

  from Science News & Analysis Agencies Rally to Tackle Big Data John Holdren, the president's science adviser, wasn't exaggerating when he said last week that “big data is indeed a big deal.” Aboutfrom Science News & Analysis Agencies Rally to Tackle Big Data John Holdren, the president's science adviser, wasn't exaggerating when he said last week that “big data is indeed a big deal.” About 1.2 zettabytes (1021) of electronic data are generated each year...

  http://www.sciencemag.org/content/336/6077/22.full
• Sreeraj m.s

  Can any one help me to combine two weak classifiers using matlab?

  Can any one help me to combine two weak classifiers using matlab, or can we combine k-NN and TREE using adaboost m2?
• G Uday Kiran

  Can anyone help me to trace the role of data mining & machine learning in Intelligent Traveler Management Systems (ITIS)?

  ITIS provide the drivers with real time travel and traffic information, such as transit routes, schedules, navigation directions and information about the delays due to congestion etc. I would likeITIS provide the drivers with real time travel and traffic information, such as transit routes, schedules, navigation directions and information about the delays due to congestion etc. I would like to know what the role of data mining and machine learning is in this regard. Can you recommend any literature?

  Recent replies ⋅ Show All (3)

  • 
    G Uday Kiran replied

    Thank U Varon... My email id id udaykiran.goru@gmail.com. Please send the articles which u have mentioned. I'll be in constant touch with u regarding this issue and please help me in this regard

    Mar 24, 2012
• Nerea Zubiaurre

  Kmeans in matlab

  I need some help with kmeans. I am using it for color based image segmentation. The problem is I want to select the cluster using some values. I mean, I prepare an image in XYZ color space forI need some help with kmeans. I am using it for color based image segmentation. The problem is I want to select the cluster using some values. I mean, I prepare an image in XYZ color space for segmentation and then I need to find a specific cluster. As kmeans does it random each time, without a specific criteria, I wanted to use a range of values of my XYZ image to find the cluster I am interested in. Does anyone have any suggestion on how to do this? Thanks.

  Recent replies ⋅ Show All (3)

  • 
    Edith Ohri replied

    Visualization can help analyzing only for simple data. It's different case when dealing with complex multi-dimension data where variables are entangled in unknown interrelations. In such a case I'dVisualization can help analyzing only for simple data. It's different case when dealing with complex multi-dimension data where variables are entangled in unknown interrelations. In such a case I'd trust only analytics algorithms, and use Maps for communication, i.e. to show the difference between the first natural data structure and the analytics clustered results.

    Mar 20, 2012
• Antonio Irpino

  Unsupervised and supervised classification methods for imprecise (probabilistic) data. Suggestion for real datasets to analyze?

  My team is working on several unsupervised techniques (K-means, dimensionality reduction tecniques) and supervised techniques (regression) for data that are described by density-like distributionsMy team is working on several unsupervised techniques (K-means, dimensionality reduction tecniques) and supervised techniques (regression) for data that are described by density-like distributions (imagine the income distribution of a country, or the pulse rate distribution of a person). We are developing techniques in the field of Symbolic Data Analysis. Until now we have adapted existing data from large dataset, but our aim is to work on real probabilistic-like data. A main problem in testing the proposed techniques (especially unsupervised ones) are the lack of benchmark datasets. We are thinking to set up a repository of such kind of data, i.e., a data table where each cell contains a density-like distribution. Contributions are welcome

  Recent replies ⋅ Show All (4)

  • 
    Antonio Irpino replied

    Thank you for the answer, very interesting topic. But we are working on data that are represented by (discrete or continuous) random variable, i.e. imagine the n peoples each one of them describedThank you for the answer, very interesting topic. But we are working on data that are represented by (discrete or continuous) random variable, i.e. imagine the n peoples each one of them described by the histogram of their pulse rate along a day, histogram of systolic pressure, and so on. In general, in a lot of cases, the domain of the probability distribution is treated as a set of states (like in the categorical case), we want to take (better) into consideration the domain of the probability distribution. Until now, we take large datasets, we do clustering and then we represents clusters as distributions over their features (that are numerical), after that we apply our methods of custering, classification or regression based on histogram data starting from a distance based on histograms. In my opinion, I imagine that this situation will be more and more frequent in the future, expecially in the era of BIG DATA, but until now I have not yet found a freely available dataset that contains directly the distributions. Secondly, it will be interesting to compare numerical and categorical distributions.

    Mar 15, 2012
• Hanumantharao Jalla

  Transformation technique to convert original data into perturbed data

  I need to transform original data into perturbed data to protect individual privacy then perform data mining technique on perturbed data , the knowledge equivalent to the original data knowledge

  Recent replies ⋅ Show All (2)

  • 
    Hanumantharao Jalla replied

    thanks joachim for your reply . simple additive noise(SAN) and multiplicative noise(MN) techniques are exist in literature but i want new technique similar to SAN and MN for my research if you havethanks joachim for your reply . simple additive noise(SAN) and multiplicative noise(MN) techniques are exist in literature but i want new technique similar to SAN and MN for my research if you have any idea please share with me Regards Hanumantharao

    Mar 14, 2012
• Bharati Kharade

  Data mining and web content mining

  How to data mining algorithms be implemented for web content mining?

  Recent replies ⋅ Show All (5)

  • 
    Mohd Din replied

    my institution has developed a website based on CMS (Content Management System) that provides information on herbs ( cultivation, safety, drugs etc ).... How can I apply data mining to this webmy institution has developed a website based on CMS (Content Management System) that provides information on herbs ( cultivation, safety, drugs etc ).... How can I apply data mining to this web contents? Anyone has any idea?

    Mar 3, 2012
• Somaye Khatiban

  How to crawl Product Review Data?

  I need some data collections of product reviews from various resources about the same set of products (domains) for instance Nokia mobile reviews from Amazon Nokia mobile reviews from eBay I need some data collections of product reviews from various resources about the same set of products (domains) for instance Nokia mobile reviews from Amazon Nokia mobile reviews from eBay Nokia mobile reviews from CNET . . . but I cannot find them in existed data collections. Does anyone know how can I download them without too much programming?