<?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: Intrarea romanilor in UE- studii de risc seismic (partea I)</title><link>http://sqlserver.ro/forums/thread/1520.aspx</link><pubDate>Thu, 11 Jan 2007 07:07:28 GMT</pubDate><guid isPermaLink="false">963f3a18-65dd-4aad-905c-b785cfe1d7c7:1520</guid><dc:creator>bogdan</dc:creator><slash:comments>0</slash:comments><comments>http://sqlserver.ro/forums/thread/1520.aspx</comments><wfw:commentRss>http://sqlserver.ro/forums/commentrss.aspx?SectionID=18&amp;PostID=1520</wfw:commentRss><description>&lt;P&gt;Cred ca depinde de algoritmul folosit. Pentru Naive Bayes, de exemplu, e perfect adevarat. De altfel, de aici vine parca si "Naive" -- de la presupunerea ca variabilele sunt independente. Arborii de decizie, pe de alta parte, nu cred ca sunt asa de sensibili la variabile corelate. Dupa primul split, dupa unul din predictori, in fiecare din noduri, variabilele corelate isi pierd importanta si al doilea split nu le va lua in considerare.&lt;/P&gt;
&lt;P&gt;Acum, cand folosesti un algoritm care nu e atat de intuitiv ca arborii de decizie sau Naive Bayes, probabil ca cel mai bun lucru e sa elimini unele variabilele puternic corelate cu cele pe care le pastrezi. Aici, o solutie solida, dar mai putin intuitiva&amp;nbsp;ar fi Principal Component Analysis, care iti transforma spatiul de variabile intr-un sub-spatiu cu mai putine dimensiuni, (aproape) independent.&amp;nbsp;Variabilele generate de PCA nu sunt foarte intuitive (reprezinta proiectii ale mai multor variabile din setul originar pe niste axe).&lt;/P&gt;
&lt;P&gt;O solutie simpla, rapida si intuitiva este Naive Bayes peste toate variabilele, marcandu-le pe toate atat Input cat si Predict (in terminologia SQL Server, asta ar fi "Predict", diferit de "Predict Only"). Dependency Net-ul construit pentru Naive Bayes arata foarte clar corelatii intre variabile. &lt;/P&gt;
&lt;P&gt;Pe baza cunostintelor despre date se poate selecta un set aproape independent de variabile.&lt;/P&gt;</description></item><item><title>Re: Intrarea romanilor in UE- studii de risc seismic (partea I)</title><link>http://sqlserver.ro/forums/thread/1510.aspx</link><pubDate>Wed, 10 Jan 2007 08:16:16 GMT</pubDate><guid isPermaLink="false">963f3a18-65dd-4aad-905c-b785cfe1d7c7:1510</guid><dc:creator>lmsasu</dc:creator><slash:comments>0</slash:comments><comments>http://sqlserver.ro/forums/thread/1510.aspx</comments><wfw:commentRss>http://sqlserver.ro/forums/commentrss.aspx?SectionID=18&amp;PostID=1510</wfw:commentRss><description>&lt;P&gt;Ca veni vorba de variabile corelate: in cartulia "DISCOVERING KNOWLEDGE IN DATA An Introduction to Data Mining" a lui Daniel Larose&amp;nbsp;scrie asa:&lt;/P&gt;
&lt;P&gt;"&lt;FONT face=Times-Roman size=2&gt;One should take care to avoid feeding correlated variables to one’s data mining and&lt;/P&gt;
&lt;P align=left&gt;statistical models. At best, using correlated variables will overemphasize one data&lt;/P&gt;
&lt;P align=left&gt;component; at worst, using correlated variables will cause the model to become&lt;/P&gt;
&lt;P&gt;unstable and deliver unreliable results."&lt;/P&gt;
&lt;P&gt;&amp;nbsp;&lt;/P&gt;
&lt;P&gt;Din experienta voastra practica, e adevarat?&lt;/P&gt;
&lt;P&gt;&amp;nbsp;Lucian&lt;/P&gt;&lt;/FONT&gt;</description></item><item><title>Re: Intrarea romanilor in UE- studii de risc seismic (partea I)</title><link>http://sqlserver.ro/forums/thread/1509.aspx</link><pubDate>Tue, 09 Jan 2007 21:54:22 GMT</pubDate><guid isPermaLink="false">963f3a18-65dd-4aad-905c-b785cfe1d7c7:1509</guid><dc:creator>ggciubuc</dc:creator><slash:comments>0</slash:comments><comments>http://sqlserver.ro/forums/thread/1509.aspx</comments><wfw:commentRss>http://sqlserver.ro/forums/commentrss.aspx?SectionID=18&amp;PostID=1509</wfw:commentRss><description>&lt;P&gt;O sa incerc sa rafinez mai incolo proiectul cu ideile tale , insa o sa urmeze un set de articole-comentarii vis-a-vis de aceste rezultate.&lt;/P&gt;
&lt;P&gt;sau chiar o sa tin cont de ele cand voi face o abordare prin Office 2007.&lt;/P&gt;
&lt;P&gt;Subiectul dar si notitele de subsol nu fac decat sa dea niste&amp;nbsp;friguri cand te gandesti ce se poate intampla ...&lt;img src="http://sqlserver.ro/emoticons/emotion-2.gif" alt="Big Smile" /&gt;&lt;/P&gt;
&lt;P&gt;Ca sa anticipez un pic , Naive Bayes a descoperit ca Lipscanii si&amp;nbsp;Calea Grivitei&amp;nbsp;sunt cele mai periculoase zone !.&lt;/P&gt;
&lt;P&gt;Rog cititorii sa nu ma transforme intr-un Hancu sau cum il chema pe celebrul prezicator de cutremure de pe la televiziuni de acu' cativa ani!&lt;img src="http://sqlserver.ro/emoticons/emotion-1.gif" alt="Smile" /&gt;&lt;/P&gt;
&lt;P&gt;Se pare ca colegul meu a obtinut datele publicate pe site-ul &lt;A href="http://www.capital.ro/"&gt;www.capital.ro&lt;/A&gt;&amp;nbsp;si cum nu-l stiu eu drept "hacker" cred ca le avea de undeva de unde &lt;/P&gt;
&lt;P&gt;se putea servi toata lumea. Asa ca n-am ce face si reatasez notitza :&lt;/P&gt;
&lt;P&gt;&lt;FONT face=Verdana&gt;"Aceste&amp;nbsp;lucruri reprezinta opinii personale ale autorului, si trebuie tratate ca atare si&amp;nbsp;fara alte implicatii."&lt;/FONT&gt;&lt;/P&gt;
&lt;P&gt;&amp;nbsp;&lt;/P&gt;
&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description></item><item><title>Re: Intrarea romanilor in UE- studii de risc seismic (partea I)</title><link>http://sqlserver.ro/forums/thread/1508.aspx</link><pubDate>Tue, 09 Jan 2007 20:32:27 GMT</pubDate><guid isPermaLink="false">963f3a18-65dd-4aad-905c-b785cfe1d7c7:1508</guid><dc:creator>bogdan</dc:creator><slash:comments>0</slash:comments><comments>http://sqlserver.ro/forums/thread/1508.aspx</comments><wfw:commentRss>http://sqlserver.ro/forums/commentrss.aspx?SectionID=18&amp;PostID=1508</wfw:commentRss><description>&lt;P&gt;&lt;FONT face=Verdana&gt;Foarte interesant! Datele sunt publice?&lt;/FONT&gt;&lt;/P&gt;
&lt;P&gt;&lt;FONT face=Verdana&gt;As incerca cateva lucruri:&lt;/FONT&gt;&lt;/P&gt;
&lt;P&gt;&lt;FONT face=Verdana&gt;-&amp;nbsp;separarea coloanei Regim Inaltime in 2 sau 3 coloane&lt;/FONT&gt;&lt;/P&gt;
&lt;UL&gt;
&lt;LI&gt;&lt;FONT face=Verdana&gt;&amp;nbsp;&amp;nbsp; - subsol (boolean)&lt;/FONT&gt;&lt;/LI&gt;
&lt;LI&gt;&lt;FONT face=Verdana&gt;&amp;nbsp; - mansarda (boolean)&lt;/FONT&gt;&lt;/LI&gt;
&lt;LI&gt;&lt;FONT face=Verdana&gt;&amp;nbsp; - numar etaje (int)&lt;BR&gt;Presupun ca toate cladirile au P(arter). O corelatie ar putea apare intre numarul de etaje si risc, corelatie ascunsa de eticheta acum&lt;/FONT&gt;&lt;/LI&gt;&lt;/UL&gt;
&lt;P&gt;&lt;FONT face=Verdana&gt;- E posibil ca expertul sa fie corelat puternic cu sectorul. E posibil ca acelasi expert sa analizeze mai multe cladiri dintr-un sector, caz in care expertiza nu mai influenteaza puternic riscul. Zona devine un factor mai important (probabil pentru ca zona + anul constructiei definesc in general in mare masura tipul constructiei). Faptul ca Sector apare in Naive Bayes (care-l trateaza ca variabila independenta) dar nu in Decision Trees (care are o maniera greedy de selectare a predictorilor), ma face sa cred ca Sector a puterninc corelat cu alta variabila. In plus, asta ar anonimiza rezultatele analizei (si&amp;nbsp;reduce riscul de litigatie &lt;img src="http://sqlserver.ro/emoticons/emotion-2.gif" alt="Big Smile" /&gt; )&lt;/FONT&gt;&lt;/P&gt;
&lt;P&gt;&lt;FONT face=Verdana&gt;Ca sa raman in zona "safe", atasez si eu nota ta finala:&lt;/FONT&gt;&lt;FONT face=Verdana&gt; "Acest mesaj reprezinta opinii personale ale autorului, si trebuie tratat ca atare si nu are alte implicatii."&lt;/FONT&gt;&lt;/P&gt;</description></item><item><title>Intrarea romanilor in UE- studii de risc seismic (partea I)</title><link>http://sqlserver.ro/forums/thread/1506.aspx</link><pubDate>Tue, 09 Jan 2007 09:42:27 GMT</pubDate><guid isPermaLink="false">963f3a18-65dd-4aad-905c-b785cfe1d7c7:1506</guid><dc:creator>ggciubuc</dc:creator><slash:comments>0</slash:comments><comments>http://sqlserver.ro/forums/thread/1506.aspx</comments><wfw:commentRss>http://sqlserver.ro/forums/commentrss.aspx?SectionID=18&amp;PostID=1506</wfw:commentRss><description>&lt;p class="MsoNormal" style="text-align:justify;"&gt;

&lt;/p&gt;&lt;p class="MsoNormal" style="text-align:justify;"&gt;Intrarea romanilor in UE- studii
de risc seismic (partea I)&lt;/p&gt;









&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&lt;/o:p&gt;&lt;span&gt;&lt;/span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; Nu va
speriati, titlul nu se refera la niste consideratii legate de vreun cutremur
provocat de intrarea&lt;br&gt;Romaniei in UE insa ce se poate spune este ca odata cu tara au intrat in UE si cladirile cu risc seismic ramase&lt;br&gt;prin cetatea lui Bucur, plasate in diverse sectoare ale capitalei.&lt;o:p&gt;&lt;/o:p&gt;&lt;br&gt;&lt;span&gt;&lt;/span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; Mi-a
atras atentia un coleg al meu preocupat de businessul asigurarilor dar si de
sperietura unui iminent&lt;br&gt;si atat de “asteptat” cutremur de pamant; acesta a
facut rost de lista de imobile, publica de altfel, care grupeaza&lt;br&gt;cladirile in 4
grupe: de la III (3) , in care respiri mai cu usurare pana la I+ (1+) unde nu
numai ca respiri greu&lt;br&gt;dar e bine sa ai niste lumanari pe aproape in caz ca …&lt;o:p&gt;&lt;/o:p&gt;&lt;br&gt;&lt;span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;Bun, gata
cu sperietura, mi-am propus sa studiez in acest articol legaturile dintre date
prin tehnicile de&lt;br&gt;data mining (minerit de date cum ne mai place sa le zicem)
si, de ce nu , sa fac niste interpretari asa cum&lt;br&gt;o sugereaza datele. As vrea sa
precizez (asa cum voi face la sfarsit), ca aceste consideratii sunt niste
opinii personale&lt;br&gt;care nu implica in nici un fel institutii, sisteme sau&lt;span&gt;&amp;nbsp; &lt;/span&gt;persoane etc. Revenind, intr-un articol
viitor voi studia aceeasi&lt;br&gt;problema prin tehnicile data mining de pe langa
Office 2007 precum si posibile noi interpretari ale datelor.&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;



&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.5in;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt; &lt;/o:p&gt;&lt;span&gt;I.&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Configurarea
studiului&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;



&lt;p class="MsoNormal" style="margin-left:0.25in;text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Iata cum arata in niste selectiuni
datele brute aflate in 4 documente pdf:&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;img src="http://sqlserver.ro/photos/snapshots/images/1498/original.aspx" height="273" width="898"&gt;&lt;br&gt;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;br&gt;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;img src="http://sqlserver.ro/photos/snapshots/images/1499/original.aspx" height="348" width="900"&gt;&lt;br&gt;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;br&gt;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.25in;text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&lt;img src="http://sqlserver.ro/photos/snapshots/images/1500/original.aspx" height="310" width="900"&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;&lt;p class="MsoNormal" style="margin-left:0.25in;text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&lt;img src="http://sqlserver.ro/photos/snapshots/images/1501/original.aspx"&gt;&amp;nbsp;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;Asa cum
observati, datele din cele 4 tabele trebuie adunate si prelucrate asa cum scrie
la carte pentru a fi o &lt;/span&gt;&lt;/p&gt;&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;sursa de date coerente si sigure. Aduse in Excel am
facute mai multe prelucrari:&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Symbol;"&gt;&lt;span&gt;·&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Am
adaugat o coloana care detine gradul de risc;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Symbol;"&gt;&lt;span&gt;·&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Am
pastrat un singur an de construire in cateva situatii (probabil acel imobil a
fost refacut in trecut) – asta nu&lt;/span&gt;&lt;/p&gt;&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt; afecteaza coerenta studiului nostru;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Symbol;"&gt;&lt;span&gt;·&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Am
adaugat campul “Expert” care desemneaza institutia sau persoana care a
expertizat imobilul; pentru imobilele &lt;/span&gt;&lt;/p&gt;&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;de grad mai scazut de risc am considerat
ca expertul a fost Primaria (Capitalei);&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Symbol;"&gt;&lt;span&gt;·&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Campul
“ArieDesfasurata” are valoarea 0 pentru imobilele in care nu s-a precizat asta;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Symbol;"&gt;&lt;span&gt;·&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Campul
RegimInaltime a ramas pe loc (apropos S+P+5E+Ma inseamna Subsol+Parter+5 Etaje+
Mansarda);&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Symbol;"&gt;&lt;span&gt;·&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;A
trebuit sa folosesc niste functii Excel ca sa omogenizez datele, de exemplu ,
imobilele cu gradul 1 aveau combinat &lt;/span&gt;&lt;/p&gt;&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;tipul de artera cu denumirea arterei&lt;span&gt;&amp;nbsp; &lt;/span&gt;( ex. Strada Academiei) si a trebuit sa duc
datele in campurile corespunzatoare.&lt;/span&gt;&lt;/p&gt;&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Sursa de date apare in figura urmatoare :&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;&lt;p&gt;&lt;img src="http://sqlserver.ro/photos/snapshots/images/1503/original.aspx" height="184" width="942"&gt;&lt;br&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;Mai
departe duc datele in SQL 2005 intr-o baza de date special facuta prin
mecanismul Import/Export al SSIS si aleg tipurile de date ale tabelei gazda asa
cum apare in figura urmatoare :&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt; &lt;/p&gt;&lt;p&gt;&lt;img src="http://sqlserver.ro/photos/snapshots/picture1515.aspx"&gt;&lt;img src="http://sqlserver.ro/photos/snapshots/images/1515/original.aspx" height="281" width="343"&gt;&lt;br&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;In
continuare deschid un proiect in SQL Server Business Intelligence Development
Studio . Configurez sursa de date si view-ul atasat dupa care trec , in ordine
la modelele bazate pe algoritmii corespunzatori (care sunt vizibili in figura
de mai jos)&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;

&lt;/p&gt;&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&amp;nbsp;&lt;/o:p&gt;&lt;/span&gt;&amp;nbsp;&lt;img src="http://sqlserver.ro/photos/snapshots/images/1504/original.aspx" height="222" width="928"&gt;&lt;/p&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;





&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;Trebuie
precizat ca IdConstr este ales drept cheia iar Risc este coloana de predictie;
An Constr si An Expertiza&lt;br&gt;au fost declarate “Discretized” astfel ca sistemul sa
formeze plaje de ani in care vor incadra constructiile respective.&lt;br&gt;Apoi trec la
procesarea datelor si sa vedem rezultatele&lt;o:p&gt;&lt;br&gt;&lt;/o:p&gt;&lt;br&gt;II Interpretarea datelor&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;







&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&lt;/o:p&gt;&lt;span&gt;&lt;/span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; Interpretarea
datelor o facem in mai multe parti , depinde de modul in care epuizam subiectul
dar si de eventuale ajustari ale&lt;br&gt;proiectului.&lt;o:p&gt;&lt;br&gt;&lt;/o:p&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; Aspectul “Dependecy Network” este vizibil , asa cum se
stie la mai multi algoritmi si, pentru o prima evaluaream grupat asa:&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;img src="http://sqlserver.ro/photos/snapshots/images/1505/original.aspx"&gt;&amp;nbsp;&lt;/p&gt;&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; Legatura mai stransa intre Risc si celelalte atribute
incepe de la “tare” la “slab” astfel :&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Symbol;"&gt;&lt;span&gt;·&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Expert,
An Constr, An expertiza, Regim Inalt in cazul Decision Trees;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.5in;text-align:justify;text-indent:-0.25in;"&gt;&lt;span style="font-size:8pt;font-family:Symbol;"&gt;&lt;span&gt;·&lt;span style="font-family:'Times New Roman';font-style:normal;font-variant:normal;font-weight:normal;font-size:7pt;line-height:normal;font-size-adjust:none;font-stretch:normal;"&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;
&lt;/span&gt;&lt;/span&gt;&lt;/span&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;Expert,
regimInalt, An Expertiza, An Constr, Sector, Tip Artera in cazul Naïve Bayes&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;



&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/o:p&gt;Ce se vede cu ochiul liber este ca atributul Expert este
identificat ca cel mai puternic pentru ambii algoritmi,&lt;br&gt;apoi anii de
constructie si de expertiza. Daca dinspre anul de constructie era de asteptat
sa primim acest rezultat tot&lt;br&gt;ce tine de aspectele de expertiza influenteaza
atributul de risc. Sa fie oare momentul si calitatea expertizei direct&lt;br&gt;legata
de aprecierea de risc ? Ramane sa atacam aceasta problema in articolele
viitoare; oricum aspectul&lt;br&gt;Dependency Network al algoritmului Microsoft
Association Rules nu afost tratat aici datorita complexitatii sale insa&lt;br&gt;va
asigur ca da rezultate bune.&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;



&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&amp;nbsp;&lt;/o:p&gt;&lt;br&gt;Gigi Ciubuc&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;



&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&amp;nbsp;&lt;/o:p&gt;&lt;br&gt;Nota. Acest articol reprezinta opinii personale ale
autorului, si trebuie tratat ca atare si nu are alte implicatii.&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&amp;nbsp;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;span&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;&lt;o:p&gt;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.25in;text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&amp;nbsp;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;

&lt;p class="MsoNormal" style="margin-left:0.25in;text-align:justify;"&gt;&lt;span style="font-size:8pt;font-family:Verdana;"&gt;&lt;o:p&gt;&amp;nbsp;&lt;/o:p&gt;&lt;/span&gt;&lt;/p&gt;</description></item></channel></rss>