Features

Spring is a time of increasing creative energy, and perhaps the least useful way to spend it is by writing Data Mining Haiku. Nevertheless, those KDnuggets readers with the most (or least) poetic talent are invited to participate in the second KDnuggets Data Mining Haiku competition.

Please email your Haiku to editor at kdnuggets dot com, and specify whether your name should be published or remain anonymous.

The winners will be announced in the next issue, determined by

• the number of data mining related references (good),
• deviations from 5,7,5 ideal count of syllables (bad),
• a sense of humor (good), and
• a totally subjective quality measure (?!?!).

Curse Bonferroni
The harder I look for it
The less it is real

(Nigel Clay, Australia)

Here is a more recent example to start this competition.

Data Mining Haiku for Adar

Preprocessing.

Cleaning, reducing
and ignoring outliers.
Only one case left.

(Arnon Hershkovitz, Israel).

About Data Mining "Haiku"
Traditional Japanese Haiku consist of three lines of 5, 7, and 5 syllables, respectively. However, considering the average level of poetic talent of data miners I know (myself included), the rules are relaxed here. Data mining haiku should still have 3 lines, but we allow a "standard deviation" of 2 syllables in each line length (i.e. lines can be 3-7, 5-9, and 3-7 syllables, respectively).

Here is an "acceptable" example with 4, 6, and 5 syllables in each line.

So much data
And so little knowledge
Keyboard waiting