📄 page_310.html
字号:
<HTML> <HEAD> <!--SCRIPT LANGUAGE="JavaScript" SRC="http://a1835.g.akamai.net/f/1835/276/3h/www.netlibrary.com/include/js/dictionary_library.js"></SCRIPT> <SCRIPT LANGUAGE="JavaScript"> if (!opener){document.onkeyup=parent.turnBookPage;} </SCRIPT!--> <META HTTP-EQUIV="Cache-Control" CONTENT="no-cache"> <META HTTP-EQUIV="Pragma" CONTENT="no-cache"> <META HTTP-EQUIV="Expires" CONTENT="-1"><META http-equiv="Content-Type" content="text/html; charset=windows-1252"><SCRIPT>var PrevPage="Page_309";var NextPage="Page_311";var CurPage="Page_310";var PageOrder="317";</SCRIPT> <TITLE>Document</TITLE> </HEAD> <BODY BGCOLOR="#FFFFFF"><CENTER><TABLE BORDER=0 WIDTH=100% CELLPADDING=0><TR><TD ALIGN=CENTER> <TABLE BORDER=0 CELLPADDING=2 CELLSPACING=0 WIDTH=100%> <TR> <TD ALIGN=LEFT><A HREF='Page_309.html'>Previous</A></TD> <TD ALIGN=RIGHT><A HREF='Page_311.html'>Next</A></TD> </TR> </TABLE></TD></TR><TR><TD ALIGN=LEFT><P><A NAME='JUMPDEST_Page_310'/><A NAME='{A3F}'/><TABLE BORDER=0 CELLSPACING=0 CELLPADDING=0 WIDTH='100%'><TR><TD ALIGN=RIGHT><FONT FACE='Times New Roman, Times, Serif' SIZE=2 COLOR=#FF0000>Page 310</FONT></TD></TR></TABLE><A NAME='{A40}'/><TABLE BORDER=0 CELLSPACING=0 CELLPADDING=0><TR> <TD ROWSPAN=5></TD> <TD COLSPAN=3 HEIGHT=17></TD> <TD ROWSPAN=5></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR><TD></TD> <TD><FONT FACE='Times New Roman, Times, Serif' SIZE=3><B>Clustering (Association and Sequencing)</B></FONT></TD><TD></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR> <TD COLSPAN=3 HEIGHT=1></TD></TR></TABLE><A NAME='{A41}'/><TABLE BORDER=0 CELLSPACING=0 CELLPADDING=0><TR> <TD ROWSPAN=5></TD> <TD COLSPAN=3 HEIGHT=12></TD> <TD ROWSPAN=5></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR><TD></TD> <TD><FONT FACE='Times New Roman, Times, Serif' SIZE=3>In order to discover possible associations and sequencing patterns in the data, the first analysis involves using a data mining tool incorporating a Self-Organizing Map (SOM), also known as a Kohonen neural network. This type of neural network is ideal for exploratory analyses and can be used to search for discrete clusters in the data. A k-nearest neighbor statistical algorithm can also be used to accomplish the same type of task. The purpose for doing this type of clustering analysis on your website data is to discover associations between visitor attributes, such as gender or age and the number of sales they make or the total amount of purchases they have made at your site.</FONT></TD><TD></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR> <TD COLSPAN=3 HEIGHT=1></TD></TR></TABLE><A NAME='{A42}'/><TABLE BORDER=0 CELLSPACING=0 CELLPADDING=0><TR> <TD ROWSPAN=5></TD> <TD COLSPAN=3 HEIGHT=12></TD> <TD ROWSPAN=5></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR><TD></TD> <TD><FONT FACE='Times New Roman, Times, Serif' SIZE=3>Clustering is a good e-mining start—it is an exploratory method of discovering relationships, which you may find surprising and totally unexpected. An SOM network explores all the data that you have compiled and gradually constructs spatial clusters of data. This type of analysis is commonly done by retailers in "market basket analysis," where they search for patterns or relationships between the sale of certain products for cross-selling and mixed promotion opportunities.</FONT></TD><TD></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR> <TD COLSPAN=3 HEIGHT=1></TD></TR></TABLE><A NAME='{A43}'/><TABLE BORDER=0 CELLSPACING=0 CELLPADDING=0><TR> <TD ROWSPAN=5></TD> <TD COLSPAN=3 HEIGHT=12></TD> <TD ROWSPAN=5></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR><TD></TD> <TD><FONT FACE='Times New Roman, Times, Serif' SIZE=3>This type of clustering analysis is also known as "unsupervised learning" because the neural network used to do the clustering is not instructed or provided a desired output. A SOM is instead designed to discover—on its own—specific clusters or classes within a given data set. For purposes of this example we will use an SOM to explore and discover clusters within a 10,000-sample data set. Hopefully it will find for us some cross-selling patterns and marketing opportunities. Using a data mining tool with a graphical programming interface we are able to import the comma-separated value (CSV) flat text file and partition a random sample of 50 percent of the records. We connect a table viewer to visually inspect the data and ensure the expected values are correct (Figure 9-3). Next, we connect a "type" node, which ensures the data types are correctly set prior to importing the data into the Kohonen (SOM) modeling node (Figure 9-4). The data mining tool will generate a graphical view from which the user can gradually begin to see the partitioning of the data into clusters identified by different shades of color and density, as shown in Figure 9-5.</FONT></TD><TD></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR> <TD COLSPAN=3 HEIGHT=1></TD></TR></TABLE><A NAME='{A44}'/><TABLE BORDER=0 CELLSPACING=0 CELLPADDING=0><TR> <TD ROWSPAN=5></TD> <TD COLSPAN=3 HEIGHT=12></TD> <TD ROWSPAN=5></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR><TD></TD> <TD><FONT FACE='Times New Roman, Times, Serif' SIZE=3>After some time (this can vary from a few minutes to a few hours depending on your CPU, the size of the data set, and the complexity of the settings of the network), the SOM will construct some distinct data cubes, or clusters. The parameters for determining the distinctions between these clusters can be set or allowed to default to the software setting. Next, as shown in Figure 9-6, the results of the clus-</FONT><FONT FACE='Times New Roman, Times, Serif' SIZE=3 COLOR=#FFFF00><!-- soft --></FONT></TD><TD></TD></TR><TR> <TD COLSPAN=3></TD></TR><TR> <TD COLSPAN=3 HEIGHT=1></TD></TR></TABLE><A NAME='{A45}'/></FORM></P></TD></TR></TABLE><P><FONT SIZE=0 COLOR=WHITE></CENTER><A NAME="bottom"> </A><!-- netLibrary.com Copyright Notice --> </BODY></HTML>⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -