<?xml version="1.0" encoding="UTF-8" ?>
<?xml-stylesheet type="text/xsl" href="http://sqlserver.ro/utility/FeedStylesheets/rss.xsl" media="screen"?><rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:slash="http://purl.org/rss/1.0/modules/slash/" xmlns:wfw="http://wellformedweb.org/CommentAPI/"><channel><title>Business Intelligence - Data Mining</title><link>http://sqlserver.ro/forums/18/ShowForum.aspx</link><description /><dc:language>en</dc:language><generator>CommunityServer 2.1 SP2 (Build: 61129.1)</generator><item><title>Re: Modele prea complexe in descoperirea de cunostinte</title><link>http://sqlserver.ro/forums/thread/7447.aspx</link><pubDate>Thu, 30 Jul 2009 15:55:23 GMT</pubDate><guid isPermaLink="false">963f3a18-65dd-4aad-905c-b785cfe1d7c7:7447</guid><dc:creator>ggciubuc</dc:creator><slash:comments>0</slash:comments><comments>http://sqlserver.ro/forums/thread/7447.aspx</comments><wfw:commentRss>http://sqlserver.ro/forums/commentrss.aspx?SectionID=18&amp;PostID=7447</wfw:commentRss><description>Inca nu &lt;img src="http://sqlserver.ro/emoticons/emotion-1.gif" alt="Smile" /&gt;&lt;br&gt;Cred ca pe la noi ar fi mai nimerita intrebarea, cati dintre voi ati folosit tehnicile KDD in viata de zi cu zi (productie)?&lt;br&gt;&lt;br&gt;Cat despre nuanta "model prea complex si rezultate eronate" , asa cum se stie folosirea algoritmului neuronal poate fi catalogat un "mers pe sarma" , adica trebuie sa fii atent cum il folosesti ca poti obtine rezultate eronate&lt;br&gt;</description></item><item><title>Modele prea complexe in descoperirea de cunostinte</title><link>http://sqlserver.ro/forums/thread/7438.aspx</link><pubDate>Thu, 30 Jul 2009 06:55:17 GMT</pubDate><guid isPermaLink="false">963f3a18-65dd-4aad-905c-b785cfe1d7c7:7438</guid><dc:creator>lmsasu</dc:creator><slash:comments>0</slash:comments><comments>http://sqlserver.ro/forums/thread/7438.aspx</comments><wfw:commentRss>http://sqlserver.ro/forums/commentrss.aspx?SectionID=18&amp;PostID=7438</wfw:commentRss><description>Pentru cei care se ocupa cu descoperirea de cunostinte, nu exista o vorba: "Daca torturezi suficient de mult datele, acestea va marturisi aproape orice". Afirmatia este sprijinita matematic de o teorema a lui Boferroni, care afirma ca "as one performs an increasing number of statistical tests, the likelihood of getting an erroneous significant finding (Type I error) also increases". Se stie, de exemplu, în situatia data în "Principle of data mining": "One particularly humorous example of this type of prediction was provided by Leinweber (personal communication) who achieved almost perfect prediction of annual values of the well-known Standard and Poor 500 financial index as a function of annual values from previous years for butter production, cheese production, and sheep populations in Bangladesh and the United States."&lt;br&gt;&lt;br&gt;Ati întâlnit in practica situatia atunci când se utilizeaza un model prea complex si rezultatele au fost eronate? Puteti prezinta o astfel de situatie, împreuna cu modul de abordare pe care l-ati folosit?</description></item></channel></rss>