CREATE MINING MODEL mortgage
(
[id] long key,
Edu_Status long DISCRETE,
Work_Status long DISCRETE,
age long CONTINUOUS,
asset_value long CONTINUOUS,
Net_income long CONTINUOUS,
paid_Status long DISCRETE PREDICT
)USING MICROSOFT_DECISION_TREES
i have a mining model above and i am designing a web cross ablication to see probablity of the a costumers' paid status.can you write for me dmx command for selecting paid status that have same criteria(ex;age=30 net income=2000,Edu_status=college) and when i write dxm window(select * from [paid_Status] ) it only shows "4" one row one column why ? as you see i dont know
dmx and datamining well..:)
string dxmcommand = "";
private AdomdCommand ascommand = new AdomdCommand();
private AdomdConnection asconnection=new AdomdConnection();
private AdomdDataReader asreader=new AdomdDataReader();
ascommand.CommandText = dmxcommand;
if I understand your question correctly, then the query would be something like:
SELECT Paid_Status, PredictProbability(Paid_Status)
FROM mortgage NATURAL PREDICTION JOIN
(SELECT 30 AS Age, 2000 as Net_Income ) AS T
This query will return the predicted value for Paid_Status as well as the proability for that prediction. Note that I did not include "College" as Edu_status. The reason is that the Edu_status column of the mining model is defined as "LONG DISCRETE" and "College" is a string, so it cannot be mapped to a long. You will have to convert "College" to the numeric code which I suppose it is used in describing this string, then add "<Numeric_Code_For_College> AS Edu_Status" to the query
Also, in C#, the sequence of operations should be along this line:
AdomdConnection cn = new AdomdConnection();
cn.ConnectionString = "Data Source=localhost; Initial Catalog=<Your Database>";
cn.Open(); // open the connection
AdomdCommand cmd = new AdomdCommand();
cmd.Connection = cn; // associate the command with the current connection
cmd.CommandText = dmxcommand;
AdomdDataReader rdr = cmd.ExecuteReader(); // obtain the reader from the command execution instead of creating it with new
Hope this helps
No comments:
Post a Comment