Data Mining WORK (prediction from the past marketing data)
$30-250 USD
Completed
Posted almost 11 years ago
$30-250 USD
Paid on delivery
I have some marketing data, I want to predict the dataset from the past data.
Data was modified to numeric values, 1 to 31.
You are expected to predict next data set by using Statistical Analysis like the Machine learning, SVM, etc...(free to select the method).
[Data explanation]
-the first column is order of data, and the following 6 column holds data, which is labeled characters (A to F).
-each column's data is selected between 1 to 31 (integer).
-the selection pattern is up to 31^6
[Output]
-the program or script which can predict 101,201,301,401,501,601,701 from the past data.
-the prediction set for no.724 (future data not included in the attached file)
-the consideration of the data in English (this is optional)
[Output example]
-In case : predicting 101 data using past data (1 to 100 data)
predicted output :(3,10,14,21,30),(2,7,8,23,28),(5,7,11,20,27), (8,13,24,26,29),(11,12,19,20,31)
real 101 data is (11,12,19,20,31), and the above output example include the one.
As i show in the example, you need to make the 5-set which include the answer [login to view URL]:you can use the F-data for prediction , but F-data is not necessary to show in the predicted data set.
[limitation/condition]
-you can use only the past data for predication
ex: in case you predict 101, you can use 1 to 100 data only.
in case you predict 701, you can use 1 to 700 data only.
-You are not allowed to use commercial tools, you can use only OSS tools or your program code.
because i need to confirm your algorithm by my computer.
-you also can use non-original code which the other people already made.
if you use such existing code, the reference souce is need to be clear.
-You can use any computer languages to analysis. R is most recommended.
-You don't need to make [login to view URL] command line interfaces is acceptable.
-the program or script are expected to work under Windows / Unix/ Linux.
I attached the data, so you can bid after trying some method to solve the problem. ([login to view URL] and [login to view URL] is the same data)
If you have any question, please feel free to contact me.
Best Regards,