Text Algorithms (6EAP)
Compression
Jaak Vilo 2012 fall
Jaak Vilo MTAT.03.190 Text Algorithms 1 Problem
• Compress – Text – Images, video, sound, …
• Reduce space, efficient communica on, etc… – Data deduplica on
• Exact compression/decompression • Lossy compression • Managing Gigabytes: Compressing and Indexing Documents and Images • Ian H. Wi en, Alistair Moffat, Timothy C. Bell • Hardcover: 519 pages Publisher: Morgan Kaufmann; 2nd Revised edi on edi on (11 May 1999) Language English ISBN-10: 1558605703 Links
• h p://datacompression.info/ • h p://en.wikipedia.org/wiki/Data_compression • Data Compression Debra A. Lelewer and Daniel S. Hirschberg – h p://www.ics.uci.edu/~dan/pubs/DataCompression.html
• Compression FAQ h p://www.faqs.org/faqs/compression-faq/ • Informa on Theory Primer With an Appendix on Logarithms by Tom
Schneider h p://www.lecb.ncifcrf.gov/~toms/paper/primer/ • h p://www.cbloom.com/algs/index.html Problem
• Informa on transmission • Informa on storage • The data sizes are huge and growing – fax - 1.5 x 106 bit/page – photo: 2M pixels x 24bit = 6MB – X-ray image: ~ 100 MB? – Microarray scanned image: 30-100 MB – Tissue-microarray - hundreds of images, each tens of MB – Large Hardon Collider (CERN) - The device will produce few peta (1015) bytes of stored data in a year. – TV (PAL) 2.7 · 108 bit/s – CD-sound, super-audio, DVD, ...
– Human genome – 3.2Gbase. 30x sequencing => 100Gbase + quality info (+ raw data) – 1000 genomes, all individual genomes … What it’s about?
• Elimina on of redundancy • Being able to predict… • Compression and decompression – Represent data in a more compact way – Decompression - restore original form • Lossy and lossless compression – Lossless - restore in exact copy – Lossy - restore almost the same informa on • Useful when no 100% accuracy needed • voice, image, movies, ... • Decompression is determinis c (lossy in compression phase) • Can achieve much more effec ve results Methods covered:
• Code words (Huffman coding) • Run-length encoding • Arithme c coding • Lempel-Ziv family (compress, gzip, zip, pkzip, ...) • Burrows-Wheeler family (bzip2) • Other methods, including images • Kolmogorov complexity • Search from compressed texts Model
Model Model
Compressed Data data Data Encoder Decoder • Let pS be a probability of message S • The informa on content can be represented in terms of bits
• I(S) = -log( pS ) bits • If the p=1 then the informa on content is 0 (no new informa on) – If Pr[s]=1 then I(s) = 0. – In other words, I(death)=I(taxes)=0 • I( heads or tails ) = 1 -- if the coin is fair • Entropy H(S) is the average informa on content of S
– H(S) = pS · I(S) = -pS log( pS ) bits h p://en.wikipedia.org/wiki/Informa on_entropy
• Shannon's experiments with human predictors show an informa on rate of between .6 and 1.3 bits per character, depending on the experimental setup; the PPM compression algorithm can achieve a compression ra o of 1.5 bits per character. • No compression can on average achieve be er compression than the entropy • Entropy depends on the model (or choice of symbols)
• Let M={ m1, .. mn } be a set of symbols of the model A and let p(mi) be the probability of the symbol mi
• The entropy of the model A, H(M) is -∑i=1..n p(mi) · log( p(mi) ) bits
• Let the message S = s1, .. sk, and every symbol si be in the model M. The informa on content of model A is -∑i=1..k log p (si) • Every symbol has to have a probability, otherwise it cannot be coded if it is present in the data http://prize.hutter1.net/ • The data compression world is all abuzz about Marcus Hu er’s recently announced 50,000 euro prize for record-breaking data compressors. Marcus, of the Swiss Dalle Molle Ins tute for Ar ficial Intelligence, apparently in cahoots with Florida compression maven Ma Mahoney, is offering cash prizes for what amounts to the most impressive ability to compress 100 MBytes of Wikipedia data. (Note that nobody is going to win exactly 50,000 euros - the prize amount is prorated based on how well you beat the current record.) • This prize differs considerably from my Million Digit Challenge, which is really nothing more than an a empt to silence people foolishly claiming to be able to compress random data. Marcus is instead looking for the most effec ve way to reproduce the Wiki data, and he’s pu ng up real money as an incen ve. The benchmark that contestants need to beat is that set by Ma Mahoney’s paq8f , the current record holder at 18.3 MB. (Alexander Ratushnyak’s submission of a paq variant looks to clock in at a dy 17.6 MB, and should soon be confirmed as the new standard.) • So why is an AI guy inser ng himself into the world of compression? Well, Marcus realizes that good data compression is all about modeling the data. The be er you understand the data stream, the be er you can predict the incoming tokens in a stream. Claude Shannon empirically found that humans could model English text with an entropy of 1.1 to 1.6 0.6 to 1.3 bits per character, which at at best should mean that 100 MB of Wikipedia data could be reduced to 13.75 7.5 MB, with an upper bound of perhaps 20 16.25 MB. The theory is that reaching that 7.5 MB range is going to take such a good understanding of the data stream that it will amount to a demonstra on of Ar ficial Intelligence. h p://marknelson.us/2006/08/24/the-hu er-prize/#comment-293
Model
Model Model
Compressed Data data Data Encoder Decoder Sta c or adap ve
• Sta c model does not change during the compression • Adap ve model can be updated during the process • Symbols not in message cannot have 0-probability • Semi-adap ve model works in 2 stages, off-line. • First create the code table, then encode the message with the code table How to compare compression techniques? • Ra o (t/p) t: original message length • p: compressed message length • In texts - bits per symbol • The me and memory used for compression • The me and memory used for decompression • error tolerance (e.g. self-correc ng code) Shorter code words…
• S = 'aa bbb cccc ddddd eeeeee fffffffgggggggg' • Alphabet of 8 • Length = 40 symbols • Equal length codewords
• 3-bit a 000 b 001 c 010 d 011 e 100 f 101 g 110 space 110 • S compressed - 3*40 = 120 bits Run-length encoding
• h p://michael.dipperstein.com/rle/index.html • The string:
• "aaaabbcdeeeeefghhhij"
• may be replaced with
• "a4b2c1d1e5f1g1h3i1j1". • This is not shorter because 1-le er repeat takes more characters... • "a3b1cde4fgh2ij" • Now we need to know which characters are followed by run-length. • E.g. use escape symbols. • Or, use the symbol itself - if repeated, then must be followed by run- length • "aa2bb0cdee3fghh1ij" Alphabe cally ordered word-lists
resume 0resume retail 2tail retain 5n retard 4rd retire 3ire
Coding techniques
• Coding refers to techniques used to encode tokens or symbols. • Two of the best known coding algorithms are Huffman Coding and Arithme c Coding. • Coding algorithms are effec ve at compressing data when they use fewer bits for high probability symbols and more bits for low probability symbols. Variable length encoders
• How to use codes of variable length? • Decoder needs to know how long is the symbol
• Prefix-free code: no code can be a prefix of another code • Calculate the frequencies and probabili es of symbols: • S = 'aa bbb cccc ddddd eeeeee fffffffgggggggg'
freq ratio p(s) a 2 2/40 0.05 b 3 3/40 0.075 c 4 4/40 0.1 d 5 5/40 0.125 space 5 5/40 0.125 e 6 6/40 0.15 f 7 7/40 0.175 g 8 8/40 0.2
Algoritm Shannon-Fano
• Input: probabili es of symbols • Output: Codewords in prefix free coding
1. Sort symbols by frequency 2. Divide to two almost probable groups 3. First group gets prefix 0, other 1 4. Repeat recursively in each group un l 1 symbol remains Example 1 a 1/2 0 b 1/4 10 c 1/8 110 d 1/16 1110 e 1/32 11110 f 1/32 11111
Example 1
Code: a 1/2 0 b 1/4 10 c 1/8 110 d 1/16 1110 e 1/32 11110 f 1/32 11111
Shannon-Fano S = 'aa bbb cccc ddddd eeeeee fffffffgggggggg'
p(s) code g 0.2 00 0.2 0.525 f 0.175 010 0.175 0.325 0.15 e 0.15 011 1 d 0.125 100 space 0.125 101 0.475 c 0.1 110 b 0.075 1110 a 0.05 1111
Shannon-Fano
• S = 'aa bbb cccc ddddd eeeeee fffffffgggggggg' • S in compressed is 117 bits • 2*4 + 3*4 + 4*3 + 5*3 + 5*3 + 6*3 + 7*3 + 8*2 = 117 • Shannon-Fano not always op mal • Some mes 2 equal probable groups cannot be achieved • Usually be er than H+1 bits per symbol, when H is entropy. Huffman code
• Works the opposite way. • Start from least probable symbols and separate them with 0 and 1 (sufix) • Add probabili es to form a "new symbol" with the new probability • Prepend new bits in front of old ones. Char Freq Code space 7 111 Huffman example a 4 010 e 4 000 f 3 1101 h 2 1010 i 2 1000 m 2 0111 n 2 0010 s 2 1011 t 2 0110 l 1 11001 o 1 00110 p 1 10011 r 1 11000 u 1 00111 "this is an example of a huffman tree" x 1 10010 • Huffman coding is op mal when the frequencies of input characters are powers of two. Arithme c coding produces slight gains over Huffman coding, but in prac ce these gains have not been large enough to offset arithme c coding's higher computa onal complexity and patent royal es • (as of November 2001/Jul2006, IBM owns patents on the core concepts of arithme c coding in several jurisdic ons). h p://en.wikipedia.org/wiki/Arithme c_coding#US_patents_on_arithme c_coding Proper es of Huffman coding
• Error tolerance quite good • In case of the loss, adding or change of a single bit, the differences remain local to the place of the error • Error usually remains quite local (proof?) • Has been shown, the code is op mal • Can be shown the average result is H+p +0.086, where H is the entropy and p is the probability of the most probable symbol Move to Front
• Move to Front (MTF), Least recently used (LRU) • Keep a list of last k symbols of S • Code – use the code for symbol. – if in codebook, move to front – if not in codebook, move to first, remove the last • c.f. the handling of memory paging • Other heuris cs ... Arithme c (en)coding
• Arithme c coding is a method for lossless data compression. • It is a form of entropy encoding, but where other entropy encoding techniques separate the input message into its component symbols and replace each symbol with a code word, arithme c coding encodes the en re message into a single number, a frac on n where (0.0 n < 1.0). • Huffman coding is op mal for character encoding (one character-one code word) and simple to program. Arithme c coding is be er s ll, since it can allocate frac onal bits, but more complicated. • Wikipedia h p://en.wikipedia.org/wiki/Arithme c_encoding • En re message is a single floa ng point number, a frac on n where (0.0 n < 1.0). • Every symbol gets a probability based on the model • Probabili es represent non-intersec ng intervals • Every text is such an interval Let P(A)=0.1, P(B)=0.4, P(C)=0.5
A [0,0.1) AA [0 , 0.01) AB [0.01 , 0.05) AC [0.05 , 0.1 )
B [0.1,0.5) BA [0.1 , 0.14) BB [0.14 , 0.3 ) BC [0.3 , 0.5 )
C [0.5,1) CA [0.5 , 0.55 ) CB [0.55 , 0.75 ) CC [0.75 , 1 )
• Add a EOF symbol. • Problem with infinite precision arithme cs • Alterna ve - blockwise, use integer- arithme cs
• Works, if smallest pi not too small • Best ra o • Problem - the speed, and error tolerance, small change has catastrophic effect • Invented by Jorma Rissanen (then at IBM) • Arithme c coding revisited by Alistair Moffat, Radford M. Neal, Ian H. Wi en - h p://portal.acm.org/cita on.cfm?id=874788
• Models for arithme c coding - • HMM Hidden Markov Models • ... • Context methods: Abrahamson dependency model • Use the context to maximum, to predict the next symbol • PPM - Predic on by Par al Matching • Several contexts, choose best • Varia ons Dic onary based compression
• Dic onary (symbol table) , list codes • If not in disc onary, use escape • Usual heuris cs searches for longest repeat • With fixed table one can search for op mal code • With adap ve dic onary the op mal coding is NP-complete • Quite good for English language texts, for example Lempel-Ziv family, LZ, LZ-77
• Use the dic onary to memorise the previously compressed parts • LZ-77 • Sliding window of previous text; and text to be compressed
/bbaaabbbaaffacbbaa…./...abbbaabab...
• Lookahead - longest prefix that begins within the moving window, is encoded with [posi on,length] • In example, [5,6] • Fast! (Commercial so ware, e.g. PKZIP, Stacker, DoubleSpace, ) • Several alterna ve codes for same string (alterna ve substrings will match) • Lempel-Ziv compression from McGill Univ. h p://www.cs.mcgill.ca/~cs251/OldCourses/1997/topic23/ Original LZ77 • Triples [posi on,length,next char] • If output [a,b,c], advance by b+1 posi ons • For each part of the triple the nr of bits is reserved depending on window length ⎡ log(n-f) ⎤ + ⎡ log(f) ⎤ + 8 where n is window size, and f is lookahead size • Example: abbbbabbc [0,0,a] [0,0,b] [1,3,a] [3,2,c] • In example the match actually overlaps with lookahead window LZ-78
• Dic onary • Store strings from processed part of the message • Next symbol is the longest match from dic onary, that matches the text to be processed • LZ78 (Ziv and Lempel 1978) • First, dic onary is empty, with index 0 • Code [i,c] - refers to dic onary (word u at pos. i) and c is the next symbol • Add uc to dic onary • Example: ababcabc → [0,a][0,b][1,b][0,c][3,c] LZW
• Code consists of indices only! • First, dic onary has every symbol /alphabet/ • Update dic onary like LZ78 • In decoding there is a danger: See abcabca – If abc is in dic onary – add abca to dic onary – next is abca, output that code – But when decoding, a er abc it is not known that abca is in the dic onary • Solu on - if the dic onary entry is used immediately a er its crea on, the 1st and last characters have to match
• Many representa ons for the dic onary. • List, hash, sorted list, combina on, binary tree, trie, suffix tree, ... LZJ
• Coding - search for longest prefix. • Code - address of the trie node • From the root of the trie, there is a transi on on every symbol (like in LZW). • If out of memory, remove these nodes/ branches that have been used only once • In prac ce, h=6, dic onary has 8192 nodes LZFG
• Effec ve LZ method • From LZJ • Create a suffix tree for the window • Code - node address plus nr of characters from teh edge. • The internal and leaf nodes with different codes • small match directly... (?) Burrows-Wheeler
• See FAQ h p://www.faqs.org/faqs/compression-faq/part2/sec on-9.html • The method described in the original paper is really a composite of three different algorithms: – the block sor ng main engine (a lossless, very slightly expansive preprocessor), – the move-to-front coder (a byte-for-byte simple, fast, locally adap ve noncompressive coder) and – a simple sta s cal compressor (first order Huffman is men oned as a candidate) eventually doing the compression. • Of these three methods only the first two are discussed here as they are what cons tutes the heart of the algorithm. These two algorithms combined form a completely reversible (lossless) transforma on that - with typical input - skews the first order symbol distribu ons to make the data more compressible with simple methods. Intui vely speaking, the method transforms slack in the higher order probabili es of the input block (thus making them more even, whitening them) to slack in the lower order sta s cs. This effect is what is seen in the histogram of the resul ng symbol data. • Please, read the ar cle by Mark Nelson: • Data Compression with the Burrows-Wheeler Transform Mark Nelson, Dr. Dobb's Journal September, 1996. h p://marknelson.us/1996/09/01/bwt/
Burrows-Wheeler Transform (BWT)
CODE: t: hat acts like this:<13><10><1 t: hat buffer to the constructor t: hat corrupted the heap, or wo W: hat goes up must come down<13 t: hat happens, it isn't likely w: hat if you want to dynamicall t: hat indicates an error.<13><1 t: hat it removes arguments from t: hat looks like this:<13><10>< t: hat looks something like this t: hat looks something like this t: hat once I detect the mangled
Example
• Decode: errktreteoe.e
• Hint: . Is the last character, alphabe cally first… Syntac c compression
• Context Free Grammar for presen ng the syntax tree • Usually for source code • Assump on - program is syntac cally correct • Comments • Features, constants - group by group Image compression
• Many images, photos, sound, video, ... Fax group 3
• Fax/group 3 • Black/white, 0/1 code • Run-length: 000111001000 → 3,3,2,1,3 • Variable-length codes for represen ng run- lengths. • Joint Photographic Experts Group JPEG 2000 h p://www.jpeg.org/jpeg2000/ • Image Compression -- JPEG from W.B. Pennebaker, J.L. Mitchell, "The JPEG S ll Image Data Compression Standard", Van Nostrand Reinhold, 1993. • Color image, 8 or 12 bits per pixel per color. • Four modes Sequen al Mode • Lossless Mode • Progressive Mode • Hierarchical Mode • DCT (Discrete Cosine Transform) • from h p://www.utdallas.edu/~aria/mcl/post/ • Lossy signal compression works on the basis of transmi ng the "important" signal content, while omi ng other parts (Quan za on). To perform this quan za on effec vely, a linear de-correla ng transform is applied to the signal prior to quan za on. All exis ng image and video coding standards use this approach. The most commonly used transform is the Discrete Cosine Transform (DCT) used in JPEG, MPEG-1, MPEG-2, H.261 and H.263 and its descendants. For a detailed discussion of the theory behind quan za on and jus fica on of the usage of linear transforms, see reference [1] below. • A brief overview of JPEG compression is as follows. The JPEG encoder par ons the image into 8x8 blocks of pixels. To each of these blocks it applies a 2- dimensional DCT. The transform matrix is normalized (element-wise) by a 8x8 quan za on matrix and then rounded to the nearest integer. This opera on is equivalent to applying different uniform quan zers to different frequency bands of the image. The high-frequency image content can be quan zed more coarsely than the low-frequency content, due to two factors. • L9_Compression/lena/ Lena Vector quan za on
• Vector quan za on • Dic onary-meetod • 2-dimensional blocks Discrete cosine transform
• A discrete cosine transform (DCT) expresses a sequence of finitely many data points in terms of a sum of cosine func ons oscilla ng at different frequencies. DCTs are important to numerous applica ons in science and engineering, from lossy compression of audio and images (where small high-frequency components can be discarded), to spectral methods for the numerical solu on of par al differen al equa ons. The use of cosine rather than sine func ons is cri cal in these applica ons: for compression, it turns out that cosine func ons are much more efficient (as explained below, fewer are needed to approximate a typical signal), whereas for differen al equa ons the cosines express a par cular choice of boundary condi ons.
• h p://en.wikipedia.org/wiki/Discrete_cosine_transform 2d DCT (type II) compared to the DFT. For both transforms, there is the magnitude of the spectrum on left and the histogram on right; both spectra are cropped to 1/4, to zoom the behaviour in the lower frequencies. The DCT concentrates most of the power on the lower frequencies. • In par cular, a DCT is a Fourier-related transform similar to the discrete Fourier transform (DFT), but using only real numbers. DCTs are equivalent to DFTs of roughly twice the length, opera ng on real data with even symmetry (since the Fourier transform of a real and even func on is real and even), where in some variants the input and/or output data are shi ed by half a sample. There are eight standard DCT variants, of which four are common.
• Digital Image Processing: h p://www-ee.uta.edu/dip/ Block Diagram of JPEG Baseline From Wallace’s JPEG tutorial (1993) JPEG tutorial Wallace’s From
ENEE631 Digital Image Lec13 – Transf. Coding & JPEG [65] Processing (Spring'04) s EE330 (Princeton) (Princeton) s EE330 ’ From Liu 475 x 330 x 3 = 157 KB luminance ENEE631 Digital Image Lec13 – Transf. Coding & JPEG [66] Processing (Spring'04) RGB Components s EE330 (Princeton) (Princeton) s EE330 ’ From Liu
ENEE631 Digital Image Lec13 – Transf. Coding & JPEG [67] Processing (Spring'04) Y U V (Y Cb Cr) Components s EE330 (Princeton) (Princeton) s EE330 ’ From Liu Assign more bits to Y, less bits to Cb and Cr ENEE631 Digital Image Lec13 – Transf. Coding & JPEG [68] Processing (Spring'04) JPEG Compression (Q=75%) s EE330 (Princeton) (Princeton) s EE330 ’
From Liu 45 KB, compression ration ~ 4:1
ENEE631 Digital Image Lec13 – Transf. Coding & JPEG [69] Processing (Spring'04) JPEG Compression (Q=75%) s EE330 (Princeton) (Princeton) s EE330 ’
From Liu 45 KB, compression ration ~ 4:1
ENEE631 Digital Image Lec13 – Transf. Coding & JPEG [70] Processing (Spring'04) JPEG Compression (Q=75% & 30%) s EE330 (Princeton) (Princeton) s EE330 ’
From Liu 45 KB 22 KB
ENEE631 Digital Image Lec13 – Transf. Coding & JPEG [71] Processing (Spring'04) Uncompressed (100KB)
JPEG 75% (18KB)
UMCP ENEE408G Slides (created by M.Wu & R.Liu2002)& © ENEE408G Slides (created byM.Wu UMCP JPEG 50% (12KB)
JPEG 30% (9KB)
JPEG 10% (5KB) ENEE631 Digital Image Lec13 – Transf. Coding & JPEG [72] Processing (Spring'04) 1.4-billion-pixel digital camera
• Monday, November 24, 2008 h p://www.technologyreview.com/compu ng/21705/page1/ • Giant Camera Tracks Asteroids • The camera will offer sharper, broader views of the sky. • The focal plane of each camera contains an almost complete 64 x 64 array of CCD devices, each containing approximately 600 x 600 pixels, for a total of about 1.4 gigapixels. The CCDs themselves employ the innova ve technology called "orthogonal transfer", which is described below.
Diagram showing how an OTA chip is made up of 64 OTCCD cells, each of which has 600 x 600 pixels Fractal compression
• Fractal Compression group at Waterloo h p://links.uwaterloo.ca/fractals.home.html • A "Hitchhiker's Guide to Fractal Compression" For Beginners p://links.uwaterloo.ca/pub/Fractals/Papers/ Waterloo/vr95.pdf • Encode using fractals. • Search for regions that with a simple transforma on can be similar to each other. • Compressin ra o 20-80 • Moving Pictures Experts Group ( h p://www.chiariglione.org/mpeg/ )
• MPEG Compression : h p://www.cs.cf.ac.uk/Dave/Mul media/node255.html • Screen divided into 256 blocks, where the changes and movements are tracked • Only differences from previous frame are shown • Compression ra o 50-100 • When one compresses files, can we s ll use fast search techniques without decompressing frst? • Some mes, yes • e.g. Udi Manber has developed a method • Approximate Matching of Run-Length Compressed Strings Veli Mäkinen, Gonzalo Navarro, Esko Ukkonen
• We focus on the problem of approximate matching of strings that have been compressed using run-length encoding. Previous studies have concentrated on the problem of compu ng the longest common subsequence (LCS) between two strings of length m and n , compressed to m' and n' runs. We extend an exis ng algorithm for the LCS to the Levenshtein distance achieving O(m'n+n'm) complexity. Furthermore, we extend this algorithm to a weighted edit distance model, where the weights of the three basic edit opera ons can be chosen arbitrarily. This approach also gives an algorithm for approximate searching of a pa ern of m le ers (m' runs) in a text of n le ers (n' runs) in O(mm'n') me. Then we propose improvements for a greedy algorithm for the LCS, and conjecture that the improved algorithm has O(m'n') expected case complexity. Experimental results are provided to support the conjecture. Kolmogorov (or Algorithmic) complexity • Kolmogorov, Chai n, ... • What is the compressed version of sequence '1234567891011121314151617181920212223242526...' ? • Every symbol appears almost equally frequently, almost "random" by entropy • for i=1 to n do print i ; • Algorithmic complexity (or Kolmogorov complexity) for string S is the length of the shortest program that reproduces S, o en noted K(S) • Condi onal complexity : K(S|T). Reproduce S given T. • h p://en.wikipedia.org/wiki/Algorithmic_informa on_theory • Algorithmic informa on theory is a field of study which a empts to capture the concept of complexity using tools from theore cal computer science. The chief idea is to define the complexity (or Kolmogorov complexity) of a string as the length of the shortest program which, when run without any input, outputs that string. Strings that can be produced by short programs are considered to be not very complex. This no on is surprisingly deep and can be used to state and prove impossibility results akin to Gödel's incompleteness theorem and Turing's hal ng problem. • The field was developed by Andrey Kolmogorov and Gregory Chai n star ng in the late 1960s. Kolmogorov complexity: size of circle in bits...
Model Model
Compressed Data data Data Encoder Decoder G J Chai n
• h p://www.cs.umaine.edu/~chai n/ • Distance using K. • d(S,T) = ( K(S|T) + K(T|S) ) / ( K(S) + K(T) ) • We cannot calculate K, but we can approximate it • E.g. by compression LZ, BWT, etc • d(S,T) = ( C(ST) + C(TS) ) / ( C(S) + C(T) ) • Use of Kolmogorov Distance Iden fica on of Web Page Authorship, Topic and Domain David Parry (PPT) in OSWIR 2005, 2005 workshop on Open Source Web Informa on Retrieval • Informatsioonikaugus by Mart Sõmermaa, Fall 2003 (in Data Mining Research seminar) h p://www.egeen.ee/u/vilo/edu/2003-04/DM_seminar_2003_II/Raport/P06/main.pdf