home \| news \| discuss \| issues \| license LURE: WHY
¶ tim@menzies.us August'17 Why \| Install \| What \| Guide \| Style LURE is a set of minimal data mining tools designed with the goal of letting students "roll their sleeves up" to muck around inside data miners. The code is written in LUA since: I can use LURE as a kind of executable specification; e.g. write this code in your favorite language). There are some interesting LUA data mining tool likes e.g. TORCH. LUA code is quite readable so for students who not know how to write LUA code; LUA code are very succinct and hence easily modifiable; LUA programs are not resource intensive; LUA code ports very easily to other platforms. LUA is fun! This code offers baseline implementations of a set of software science operators which, I say, are needed for data science. But you should be very critical of the technical choices I made in that implementation. What simplifications did I make? What better technologies should I use? What did I overlook? And (here's the trap) if you think you can handle the above in (e.g.) TensorFlow or Torch or using 100 other methods, I would lean forward and say "yes? really? show me how". Status LURE is about about one-third built and about one-tenth tested. But its good to have dreams since "a man's (sic) reach should exceed his grasp, Or what's a heaven for?". Motivation There has been much recent work applying data miners to software engineering. But what about the other way around? What software engineering principles should be apply to data miners? After decades of use of data miners, what services should we demand from our data miners, and how do we build those services? The premise of LURE is that data miners are not black boxes that we should buy, then uncritically use "as-is", without modification. Rather, data miners are just software and software contains 100s of design choices that effects what is learned: Some of those choices may not be appropriate for your specific needs. Also, some of those choices might actually be sub-optimum for your domain. Further, if you actually understand the internals of a data miner, you can mix and match what data miners do in order to provide useful and innovative solutions to your data mining tasks.	require "show"
¶ Legal LURE, Copyright (c) 2017, Tim Menzies All rights reserved, BSD 3-Clause License Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

home | news | discuss | issues | license

LURE: WHY

tim@menzies.us
August'17

LURE is a set of minimal data mining tools designed with the goal of letting students "roll their sleeves up" to muck around inside data miners.

The code is written in LUA since:

I can use LURE as a kind of executable specification; e.g. write this code in your favorite language).
There are some interesting LUA data mining tool likes e.g. TORCH.
LUA code is quite readable so for students who not know how to write LUA code;
LUA code are very succinct and hence easily modifiable;
LUA programs are not resource intensive;
LUA code ports very easily to other platforms.
LUA is fun!

This code offers baseline implementations of a set of software science operators which, I say, are needed for data science. But you should be very critical of the technical choices I made in that implementation. What simplifications did I make? What better technologies should I use? What did I overlook? And (here's the trap) if you think you can handle the above in (e.g.) TensorFlow or Torch or using 100 other methods, I would lean forward and say "yes? really? show me how".

Status

LURE is about about one-third built and about one-tenth tested. But its good to have dreams since "a man's (sic) reach should exceed his grasp, Or what's a heaven for?".

Motivation

There has been much recent work applying data miners to software engineering. But what about the other way around? What software engineering principles should be apply to data miners? After decades of use of data miners, what services should we demand from our data miners, and how do we build those services?

The premise of LURE is that data miners are not black boxes that we should buy, then uncritically use "as-is", without modification. Rather, data miners are just software and software contains 100s of design choices that effects what is learned:

Some of those choices may not be appropriate for your specific needs.
Also, some of those choices might actually be sub-optimum for your domain.
Further, if you actually understand the internals of a data miner, you can mix and match what data miners do in order to provide useful and innovative solutions to your data mining tasks.

require "show"

Legal

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.