-
Notifications
You must be signed in to change notification settings - Fork 8
/
ReadMe.html
455 lines (381 loc) · 20.3 KB
/
ReadMe.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<meta name="GENERATOR" content="TtH 3.59">
<style type="text/css"> div.p { margin-top: 7pt;}</style>
<style type="text/css"><!--
td div.comp { margin-top: -0.6ex; margin-bottom: -1ex;}
td div.comb { margin-top: -0.6ex; margin-bottom: -.6ex;}
td div.hrcomp { line-height: 0.9; margin-top: -0.8ex; margin-bottom: -1ex;}
td div.norm {line-height:normal;}
span.roman {font-family: serif; font-style: normal; font-weight: normal;}
span.overacc2 {position: relative; left: .8em; top: -1.2ex;}
span.overacc1 {position: relative; left: .6em; top: -1.2ex;} --></style>
<title> ``Read Me'' for \CSSR </title>
<h1 align="center">"Read Me" for <tt>CSSR</tt> </h1>
<h3 align="center">Kristina Lisa Klinkner and Cosma Rohilla Shalizi<br />
<rot13 [email protected]> <rot13 [email protected]></h3>
<h3 align="center">Last revised 11 September 2005 </h3>
<div class="p"><!----></div>
<h1>Contents </h1><a href="#tth_sEc1"
>1 General</a><br />
<a href="#tth_sEc2"
>2 Obtaining and Installing the Program</a><br />
<a href="#tth_sEc3"
>3 Usage</a><br />
<a href="#tth_sEc4"
>4 Some Suggestions About Parameters</a><br />
<a href="#tth_sEc5"
>5 Known Issues</a><br />
<a href="#tth_sEc6"
>6 Bug Reports, Fixes, Modifications</a><br />
<a href="#tth_sEc7"
>7 Some Details on the Code</a><br />
<a href="#tth_sEc7.1"
>7.1 Revision History</a><br />
<a href="#tth_sEc8"
>8 Credits and Acknowledgments</a><br />
<div class="p"><!----></div>
<h2><a name="tth_sEc1">
1</a> General</h2>
<div class="p"><!----></div>
<tt>CSSR</tt> tries to infer the minimal Markovian model capable of generating a
time-series, or set of time-series from the same source. The program
implements the algorithm proposed in the paper ""Blind Construction of Optimal
Nonlinear Recursive Predictors for Discrete Sequences", hereafter
BC.<a href="#tthFtNtAAB" name="tthFrefAAB"><sup>1</sup></a> We won't describe the algorithm
in any detail here (see BC for that), but the next two paragraphs say a little
about what it produces and how it does it.
<div class="p"><!----></div>
The output of the algorithm is a set of states which form a Markov chain. Each
state has a certain probability of emitting any of the symbols in the original
time series. The current state and the symbol it emits fix the next state.
(The states are "future-resolving", if you're from nonlinear dynamics, or
"deterministic", if you're from automata theory.) Each state, moreover,
corresponds to a distinct set of strings, in the following sense. If the state
<i>A</i> contains a string <i>w</i>, and at time <i>t</i> the time-series ends with <i>w</i>, then
at time <i>t</i> the Markov chain is in state <i>A</i>. The set of states, their
transition probabilities and connections, is called the state machine.
<div class="p"><!----></div>
The algorithm uses a recursive inference procedure to find the simplest set of
states with the above properties that can reproduce the statistical properties
of the data. If we could give the algorithm an infinitely long time series,
and let it consider infinitely long sub-strings, it would produce the
<i>causal states</i> of the process, which are its ideal predictors (see
BC for a formal definition). Since we have only finite data, there is
always some probability that the inferred or estimated states are not the true
causal states. Nonetheless, for the rest of this file, when we say "causal
states", we mean the estimated causal states.
<div class="p"><!----></div>
<h2><a name="tth_sEc2">
2</a> Obtaining and Installing the Program</h2>
<div class="p"><!----></div>
The code for <tt>CSSR</tt> can be obtained from <a href="http://bactra.org/CSSR/"><tt>http://bactra.org/CSSR/</tt></a>. If
you'd like to set up a new archive for it, we'd appreciate hearing about it.
<div class="p"><!----></div>
From any site, download the CSSR-v0.1.tar.gz file for the code. When gunzipped
and untarred, this will produce a directory called CSSR-v0.1, containing all
the necessary header and source code files, a copy of this documentation, the
release license, and a make file. Running <tt>make</tt> inside that directory
will produce an executable, which should be moved to someplace in the command
path. On most Unix systems, the following sequence of commands will create the
executable and put it in the your <tt>bin</tt> directory, usually part of your
command path.
<pre>
gunzip CSSR-v0.1.tar.gz
tar xvf CSSR-v0.1.tar
cd CSSR-v0.1
make
cp CSSR ~/bin/
</pre>
<div class="p"><!----></div> The code has been successfully compiled with gcc
3.1 on Macintosh OS X (10.2), with gcc 3.2 on Linux (Red Hat 9, SUSE 9), with
gcc 3.3 on Macintosh OS X (10.3) and with Microsoft Visual C++ on Windows 98.
On some systems, compilation may produce warnings about escape sequences or the
use of deprecated headers. These can be safely ignored.
<div class="p"><!----></div>
<h2><a name="tth_sEc3">
3</a> Usage</h2>
<a name="usage">
</a>
<div class="p"><!----></div>
<tt>CSSR</tt> is a command-line program.
<div class="p"><!----></div>
<center><tt>CSSR</tt> <i>alphabetfile</i> <i>datafile</i>
<i>maxlength</i> <tt>[-m] [-s siglevel] [-ch]</tt>
</center>
<div class="p"><!----></div>
The first argument is the name of a file which contains all the symbols in the
alphabet used by your data. The second argument is the name of the data file
itself. Only one data file can be used, but it may contain multiple lines.
The third argument is the maximum length of history to examine, which we will
abbreviate by <i>L</i> here. The three trailing flags are optional arguments. Set
the <tt>-m</tt> flag if the data-file contains multiple lines (see below). If
the <tt>-s</tt> flag is set, it must be followed by a significance level; the
default value is 0.001. (For more on setting parameters, see Section
<a href="#parameters">4</a>.) If the <tt>-ch</tt> flag is set, the program will use
the <font face="symbol">c</font
><sup>2</sup> significance test, rather than the default Kolmogorov-Smirnov
test.
<div class="p"><!----></div>
In multiple-line mode (entered through the <tt>-m</tt> flag), each line of the
data file is treated as a distinct time series from the same source.
(Technically, as independent realizations of a single stochastic process.) The
lines need not be of equal length.
<div class="p"><!----></div>
The program will create the following files after running, where
<i>datafile</i> is the name of the file the data is in:
<ol type="1">
<li> <i>datafile</i>_results
<div class="p"><!----></div>
</li>
<li> <i>datafile</i>_info
<div class="p"><!----></div>
</li>
<li> <i>datafile</i>_state_series
<div class="p"><!----></div>
</li>
<li> <i>datafile</i>_inf.dot
<div class="p"><!----></div>
</li>
</ol>
<div class="p"><!----></div>
(1) contains the information on the inferred states and their properties. For
each state, it gives:
<ul>
<li> the histories of length <i>L</i><font face="symbol">-</font
>1 and <i>L</i> in the state
<div class="p"><!----></div>
</li>
<li> the probability that the state will emit different symbols (e.g. <i>P</i>(<i>a</i>) = <i>x</i>) and the states transitioned to when those symbols are emitted (e.g. <i>T</i>(<i>a</i>) = <i>s</i>)
<div class="p"><!----></div>
</li>
<li> the observed probability of the state in the data-stream
<div class="p"><!----></div>
</li>
</ul>
<div class="p"><!----></div>
(2) is the file containing the metrics run on the causal state machine. These
are:
<ul>
<li> the number of states
<div class="p"><!----></div>
</li>
<li> the statistical complexity (entropy of the states)
<div class="p"><!----></div>
</li>
<li> the entropy rate
<div class="p"><!----></div>
</li>
<li> three measures of the difference between the empirical distribution of
symbol sequences, and that generated by the inferred causal state machine.
These are: the divergence or relative entropy between the inferred and
empirical distribution; the relative entropy rate, or increase per symbol in
the divergence; the total variational distance ("variation") between the
two distributions.
<div class="p"><!----></div>
</li>
</ul>
<div class="p"><!----></div>
Note that the relative entropy and the relative entropy rate can be infinite;
this indicates that the inferred model gives a probability of zero to a
sequence in the data.
<div class="p"><!----></div>
Note that sometimes, when the relative entropy should be very small (order of
10<sup><font face="symbol">-</font
>6</sup> bits or less), numerical rounding errors result in a negative
number being calculated. In these cases, the program outputs zero. Similarly,
the complexity of one-state machines is sometimes reported as <font face="symbol">-</font
>0.
<div class="p"><!----></div>
(3) is the series of causal states in the data. That is, the program scans
through the data, looks up which state the history-to-date is in, and writes
the corresponding symbol to this file. What you see is then the trajectory
through estimated causal state space of the data/process. Multiline data
results in a multiline state-series file.
<div class="p"><!----></div>
(4) represents the state machine as a labeled directed graph, where each state
has its own node, and each transition between states its own edge. The file is
for use with the program <tt>dot</tt>, available from
<a href="http://www.graphviz.org/"><tt>http://www.graphviz.org/</tt></a>.
<div class="p"><!----></div>
<h2><a name="tth_sEc4">
4</a> Some Suggestions About Parameters</h2>
<a name="parameters">
</a>
<div class="p"><!----></div>
It is always good to use as much data as you can. While it is generally good
practice to hold back some data for testing or cross-validation, we recommend
that this be minimized. High-entropy processes are especially data-hungry.
(See BC.) For reference, let us call the number of data-points <i>N</i>.
<div class="p"><!----></div>
The two key parameters of the program are the maximum history length, <i>L</i>, and
the significance level used in the test, <i>s</i>. For any given process, there is a
minimum history length <font face="symbol">L</font
>, such that the true states cannot be found if
<i>L</i> < <font face="symbol">L</font
>. The number of states returned may be less than the correct
number <em>or higher</em>. If <i>L</i> <font face="symbol">³</font
> <font face="symbol">L</font
>, and there is enough data, there
will generally be a "plateau" of values of <i>L</i> where the correct number of
states is returned. For fixed <i>N</i>, if we keep increasing <i>L</i>, then past a
certain point there are not enough examples of each string in the data. This
tends to erroneously create new states, which spawn others through
determinization. Thus there is generally a "blow-up" when <i>L</i> is too large
(relative to <i>N</i> and <i>s</i>). A rough guide-line is to limit <i>L</i> to no more than
log<sub>2</sub><i>N</i>/log<sub>2</sub><i>k</i>, where <i>k</i> is the alphabet size (see BC for
details).
<div class="p"><!----></div>
In general, one should use as small an <i>L</i> as possible, since under-sampling,
even before the blow-up, will reduce the accuracy of many probability
estimates. Blow-up can be delayed by reducing <i>s</i> - that is, reducing the
probability of mistakenly splitting a state - but this carries the risk of
failing to create valid new states. We suggest exploring the data at low
<i>L</i> and high <i>s</i> initially, and then increasing <i>L</i> and lowering <i>s</i>. If a
stable architecture is found, it should be recorded at the lowest possible <i>L</i>.
<div class="p"><!----></div>
<h2><a name="tth_sEc5">
5</a> Known Issues</h2>
<div class="p"><!----></div>
There are a few known issues with <tt>CSSR</tt> 's behavior. These arise from certain
unavoidable approximations in the determinization procedure. (For details, see
below.)
<div class="p"><!----></div>
After creating and determinizing the states, <tt>CSSR</tt> goes over the input data and
attempts to produce the corresponding sequence of causal states, both as a
filtering procedure, and as part of estimating the fit of the causal-state
model (see Section <a href="#usage">3</a>, item (3)). Typically there will be some initial
number of symbols which it must read before "synchronizing" to one of the
causal states, after which it will follow the transitions deterministically,
and never de-synchronize. Because of the approximations mentioned, it can
happen that certain transitions are deleted from the causal state model which
shouldn't be. This can lead to three kinds of problem: (1) <tt>CSSR</tt> can never
synchronize to a state at all; (2) it has to re-synchronize at some point; (3)
it encounters an apparently "impossible" sequence in the data.
<div class="p"><!----></div>
In case (1), <tt>CSSR</tt> writes the message "Error: Could not synchronize, cannot
use this data" to standard error and halts execution. In case (2), it
produces a warning message on both standard error and the <tt>_info</tt>
output file. In case (3), it produces a warning, and discounts that particular
string from various calculations.
<div class="p"><!----></div>
The best approach to these problems is to use a longer history length, if
possible, and to provide more data. In the case of a third error, it can
sometimes arise if a particular string appears only once, at the very beginning
of the data, and sometimes removing that string from the data-file fixes
matters.
<div class="p"><!----></div>
All three errors arise because we have only finite-length histories available
to us, while what we want are really infinite ones. This forces us to make
certain approximations in our implementation of the theory. Specifically, in
the determinization procedure, we are forced to "guess" which state certain
strings belong to, even though we have not directly examined these strings.
The particular approximation scheme (or "closure") we have adopted may be
investigated by consulting the code in AllStates.cpp. (Others are possible, but
this one seemed to give the best over-all results.) Sometimes this closure will
"guess wrong", and possible transitions will be labeled forbidden, etc. In
these cases, extending the history length <em>should</em> solve the problem, if
enough data is available for reliable inference. Similar approximations must
be made in determining whether or not a given state is transient or recurrent
on the basis of finite data. This occasionally leads to a recurrent state being
labeled transient, which in turn is the most common cause of the code mistaking
an actually-occurring string for an impossible one. Again, the best approach
is to provide more data, and a longer history.
<div class="p"><!----></div>
<h2><a name="tth_sEc6">
6</a> Bug Reports, Fixes, Modifications</h2>
<div class="p"><!----></div>
We welcome bug reports or reports of strange behavior. These reports are
welcomed with more enthusiasm when accompanied by successful modifications to
the code! (See the accompanying file on the Gnu Public License for information
about modifying the code.) Even if you can't fix it, however, please do tell
us about it; at the least it will be documented for other users.
<div class="p"><!----></div>
If you modify <tt>CSSR</tt> , and want to make the resulting program available, please
let us know. We are happy to provide a link, and have a (limited) capability
to host alternate versions and descendants. Also, if you use <tt>CSSR</tt>
successfully in some application, we'd love to hear about it.
<div class="p"><!----></div>
Please check <a href="http://bactra.org/CSSR/"><tt>http://bactra.org/CSSR/</tt></a> for up-to-date contact
information.
<div class="p"><!----></div>
<h2><a name="tth_sEc7">
7</a> Some Details on the Code</h2>
There are twelve classes in the program. They are each comprised
of a .cpp file and a .h file, except for ArrayElem (contained in
G_Array) and StringElem (contained in State), as well as a source
file Main, and header files Common.h and Main.h. <br />
<div class="p"><!----></div>
<table>
<tr><td>AllStates </td><td>Contains and manipulates growable array of states</td></tr>
<tr><td>ArrayElem </td><td>Each element in the growable array </td></tr>
<tr><td>G_Array </td><td>A generic growable array</td></tr>
<tr><td>Hash </td><td>Hash table which points from histories to their parent states</td></tr>
<tr><td>Hash2 </td><td>Hash table which points from indices to symbol/alpha values</td></tr>
<tr><td>Machine </td><td>Manipulates the determinized state machine and runs metrics</td></tr>
<tr><td>ParseTree </td><td>Reads in data file and stores all strings present in file</td></tr>
<tr><td></td><td>up to length <i>L</i> the maximum length input at the command line</td></tr>
<tr><td>States </td><td>A state, contains all data for a single state</td></tr>
<tr><td>StringElem </td><td>the element containing the history for a single state</td></tr>
<tr><td>Test </td><td>Performs statistical significance tests</td></tr>
<tr><td>TransTable </td><td>Stores initially estimated transitions from all
histories of length </td></tr>
<tr><td></td><td><i>L</i> in any given state. This class is
used by AllStates to check for</td></tr>
<tr><td></td><td>transient states before
determinization.
</td></tr></table>
<div class="p"><!----></div>
For brief descriptions of the classes, see the top of their source files. Note
that the terms "string" and "history" are used interchangeably in the
program. The terms both correspond to the concept of a "history" (as
described in BC), but the program implements these as strings of symbols.
<div class="p"><!----></div>
Also, after initial state splitting, all strings/histories of less than maximum
length minus one are deleted. This has no effect on the outcome of the
algorithm and saves time and space.
<div class="p"><!----></div>
Lastly, the removal of transient states prior to determinization implemented in
the <tt>AllStates::CheckConnComponents</tt> procedure is not strictly
necessary. With a different implementation, the deletion of these states could
automatically occur during the determinization process itself (see the
pseudocode in BC for details), and the outcome of the algorithm would be the
same. As it is implemented here the code is slightly redundant.
<div class="p"><!----></div>
<h3><a name="tth_sEc7.1">
7.1</a> Revision History</h3>
<div class="p"><!----></div>
<table>
<tr><td>0.1 </td><td>11 September 2005 </td><td>Minor bug-fixes to multi-line mode</td></tr>
<tr><td>0.1 </td><td>7 August 2005 </td><td>Corrects minor numerical bugs. These only affected</td></tr>
<tr><td></td><td></td><td>the calculations of relative entropy, relative entropy</td></tr>
<tr><td></td><td></td><td>rate and total variation distance, and then only by</td></tr>
<tr><td></td><td></td><td>about <i>O</i>(<i>k</i>/<i>N</i>).</td></tr>
<tr><td>0.0 </td><td>7 August 2003 </td><td>First public release.
</td></tr></table>
<div class="p"><!----></div>
<h2><a name="tth_sEc8">
8</a> Credits and Acknowledgments</h2>
<div class="p"><!----></div>
<tt>CSSR</tt> was written by KLK, with some help in debugging and testing from CRS;
this documentation was jointly written. Parts of the <tt>CSSR</tt> code were written
at the Santa Fe Institute and at the University of San Francisco. Support at
SFI was provided by SFI's core grants from the National Science Foundation and
the MacArthur Foundation, by the NSF's Research Experience for Undergraduates
program, and by the Dynamics of Learning group, under DARPA contractual
agreement F30602-00-2-0583. Support at USF came from the Clare Boothe Luce
Foundation.
<div class="p"><!----></div>
<hr /><h3>Footnotes:</h3>
<div class="p"><!----></div>
<a name="tthFtNtAAB"></a><a href="#tthFrefAAB"><sup>1</sup></a>Cosma Rohilla Shalizi and Kristina Lisa Shalizi, pp. 504-511 of
Max Chickering and Joseph Halpern (eds.), <em>Uncertainty in Artificial
Intelligence: Proceedings of the Twentieth Conference</em>, available from
<a href="http://arxiv.org/abs/cs.LG/0406011"><tt>http://arxiv.org/abs/cs.LG/0406011</tt></a>.
<br /><br /><hr /><small>File translated from
T<sub><font size="-1">E</font></sub>X
by <a href="http://hutchinson.belmont.ma.us/tth/">
T<sub><font size="-1">T</font></sub>H</a>,
version 3.59.<br />On 9 Aug 2005, 19:11.</small>
</html>