Temporal Data Mining
====================

Status: Alpha

The Temporal Data Mining(TDM) code is a restricted variant of association rule
data mining.  In this case, temporal data is also considered as part of the
association process.

Due to deprecation of the clCreateImage2d command, TDM is known to lack support
for OpenCL specification 1.2

Running
-------

Usage: tdm [<platform> <device> | <type> --] <filepath> <temporal constraints path> <threads> <support>
       <(a)bsolute | (r)atio> <(s)tatic |(d)ynamic>
       <(m)ap and merge | (n)aive | (h)ybrid>

    <filepath>                  :file containing temporal database
    <temporal constraints path> :path containing temporal constraints  
    <threads>                   :maximum number of threads to use      
    <support>                   :culling threshold value                       
    <(a)bsolute | (r)atio>      :culling threshold type                
    <(s)tatic | (d)ynamic>      :static or dynamic memory management
    <(m)ap and merge | (n)aive  :algorithm type to use
        | (h)ybrid>           

	<platform>                  :integer ID of platform to use
    <device>                    :integer ID of device in <platform> to use
    <type>		: device type to use (0:CPU, 1:GPU, 2:MIC, 3:FPGA)	

Example: tdm 0 0 -- test/finite-state-machine/tdm/stream-new-1.csv test/finite-state-machine/tdm/streamIntervals.txt 100 0.5 r s n