Bruce Fast
Member
Member # 924
|
posted 25. March 2006 00:24
quote:
Could you describe to me what you mean by the "complexity of a pattern". I have a concept of a pattern as either being a code (or embodied code) that has some symmetry, or some algorithm forming it.
Ok, it's time for a bit of data compression theory. If you have truly random information, and you plot it on a graph, you get a rather boring graph. However, so often when you process "real world" data, you get patterns. The most common pattern is the bell curve. The pattern isn't a pattern such as what you get when you turn 1/7 into a decimal -- a way of predicting what the next value in the string will be. Rather, the numbers at the mode of the curve happen more often than numbers on the fringe.
If we see data that renders in a bell curve, we can use that fact to compress the data. With the huffman algorithm, for instance, rather than rendering each value in, say 16 bytes of data, we render the more common (nearer the mode) values in, say 4 bytes and less common values in, say 20 bytes. Because the 4 byte units happen more often, the total length of data is, well, compressed. (Please feel free to google huffman, to get more details.)
Bottom line, though the bell curve is a "pattern", it does not give you any ability to determine what the next value in the string will be. Hence it is a pattern in complex information.
IP: Logged
|
Printer-friendly view of this topic