1 Discovering Unexpected Information from Your Competitor’s Web Sites Bing Liu, Yiming Ma, Philip...
-
date post
21-Dec-2015 -
Category
Documents
-
view
218 -
download
5
Transcript of 1 Discovering Unexpected Information from Your Competitor’s Web Sites Bing Liu, Yiming Ma, Philip...
1
Discovering Unexpected Information from Your
Competitor’s Web Sites
Bing Liu, Yiming Ma, Philip S. Yu
Héctor A. Villa Martínez
2
Objective of this article
The authors presents a system to help find unexpected information in a web site.
3
Searching information in the web
Many methods Keyword based (e.g. Goggle, Yahoo). Wrapper based (e.g. extract prices). Web query languages (e.g. extend SQL). User preference based (specify categories).
4
Searching information in the web
Main drawbacks: Hard to find unexpected information. Only finds anticipated information.
5
What is unexpected anyways?
A piece of information is unexpected if: it is relevant but unknown, or it contradicts existing beliefs or expectations
relevant interesting (subjective)
6
Summary of the approach
U: user web site
E: knowledge about the competitor
C: competitor web site
Compare C vs. U and E to find unexpected information in C.
7
How to compare two web pages
Use the vector space representation: Define a set of p keywords (index terms)
K = {k1, k2, …, kp). Represent a document D using a vector
D = {w1, w2, …, wp} where wi is the weight of the keyword i
wi > 0 if keyword i appears in D = 0 otherwise
8
Vector space representation
Example:K = {night, day, empire, barbarians, people, house}D = [“Because night is here but the barbarians have not
come.And some people arrived from the borders,and said that there are no longer any barbarians.And now what shall become of us without any barbarians?Those people were some kind of solution.”]D = {1, 0, 0, 3, 2, 0} or normalized to:D = {1/6, 0, 0, 3/6, 2/6, 0}
9
Comparing two web pages
Given two web pages in vector space representation, D = {d1, d2, …, dn}, and Q = {q1, q2, …, qn} the cosine gives a measure of similarity:
sim (D, Q) = (D ● Q) / (|D| * |Q|)
10
Comparing two web pages
Example:
P = {0.3, 0.0, 0.0, 0.7}
Q = {0.5, 0.0, 0.1, 0.4}
R = {0.0, 0.5, 0.5, 0.0} Sim (P, P) = (P ● P) / (|P| * |P|) = 1.0 Sim (P, Q) = (P ● Q) / (|P| * |Q|) = 0.87 Sim (P, R) = (P ● R) / (|P| * |R|) = 0
11
Methods to find unexpected information in a site
Let U = (u1, …, um) the user web site, and C = (c1, …, cn) the competitor web site:
1. Find the corresponding C page(s) of a U page.
2. Find unexpected terms in a C page.3. Find unexpected pages in C.4. Find unexpected concepts in a C page.5. Find unexpected outgoing links.
12
1. Find the corresponding C page(s) of a U page
Given a page ui in U
Compare ui with each page in C. Order the results in descending order.
13
1. Find the corresponding C page(s) of a U page
Example: Select u1
Find sim(u1, c1), sim(u1, c2), …, sim(u1, cn) Order the results in decreasing order: say c4,
c2, c8, … etc.Complexity:
O(G|C| + |ui||C|)
where G = max number of terms in cj
14
2. Find unexpected terms in a C page
Given uj and ci measure the unexpectedness of each term tr
1 – (frj / fri) if (frj / fri) ≤ 1
unexpTrij =
0 otherwise
15
2. Find unexpected terms in a C page
Example:
keywords = {data, predict, classify, state}
uj = {0.4, 0.5, 0.0, 0.1}
ci = {0.3, 0.3, 0.2, 0.2}
unexpT = {0, 0, 1, 0.5}
Complexity: O(Z)
where Z = number of terms in cj
16
3. Find unexpected pages in C
1. Combine all pages of U in a single page Du.
2. Combine all pages of C in a single page Dc.
3. Compute the unexpectedness of each term kt in Dc with respect to Du. (Task 2)
4. The unexpectedness of a page Ci is the sum of the unexpectedness of its terms
5. unexpPi = (ΣunexpTrcu) / m
17
3. Find unexpected pages in C
Complexity
O(Mu|U| + Mc|C|)
where
Mu is the maximal number of terms in a U page
Mc is the maximal number of terms in a C page
18
4. Find unexpected concepts in a C page
A concept is a set of keywords that occur together and express the same idea.
Example: “information extraction”, “extraction of information”, and “information is extracted” express the same idea “information extraction”
19
4. Find unexpected concepts in a C page
Algorithm Divide the page in sentences. Use the Apriori algorithm (Agrawal &
Srikant) to find association rules of the form X Y with confidence c, where X and Y K, the set of keywords and c is user defined. These association rules are the concepts present in the page.
20
4. Find unexpected concepts in a C page
3. Treating each concept as a term, proceed as Task 2, finding unexpected terms in C.
21
5. Find unexpected outgoing links
Let Lu be the set of outgoing links from U
Let Lc be the set of outgoing links from C
unexpL = Lc – Lu
22
Incorporating user knowledge
Let E be the user knowledge about his competitor. E is specified as:
Keyword terms Concepts Links
23
Incorporating user knowledge
The elements in E are incorporated in task 2 thru 5 to find unexpected terms, pages, concepts, and links.
Elements in E are ranked low in unexpectedness.
24
System architecture
C++/Win32 A spider or crawler. Collects information. Keyword extractor & concepts finder. Comparison component. Do tasks 1-5. User interface.
25
A running example
The authors compare its own site with SGI’s MineSet data mining site, and not extra knowledge:
http://www.comp.nus.edu.sg/~dm2
http://www.sgi.com/software/mineset
26
Results
Found documentation pages in SGI site. Now the authors are planning to add their own.
Found previously unknown pages describing MineSet technology.
Found some previously unknown MineSet features.
Found many interesting terms, concepts, and links.
27
Evaluation
The system was further tested with three different organizations:
Travel company Private school Diving company
28
Evaluation
The users reported the system helped them in: Focus in interesting pages, terms, and
concepts. Make a more complete analysis of the
competitor’s site. Not missing important information. Find unexpected things.
29
Efficiency
If number of keywords is constant, the algorithms are linear in the number of pages.
Tested on a Pentium II 350 PC with 64MB of RAM #pag sim unexpTrij unexpPj Assoc. mining[1] (143) 0.0128 0.0156 0.0232 0.0379[2] (21) 0.0134 0.0189 0.0213 0.0182[3] (66) 0.0113 0.0177 0.0198 0.0206[4] (127) 0.0097 0.0201 0.0224 0.0115[5] (46) 0.0143 0.0153 0.0188 0.0105
[1] http://www.bluemartini.com [4] http://www.sgi.comlsoftwarelmineset [2] http://www.datamining.com [5] http://www.spss.comlclementine [3] http://www.mineit.com
30
Future work
Use of metadata. Study how links can be used to infer more
unexpected information. Monitor the site, reporting any unexpected
change.
31
Intrinsic limitations
Text oriented. Do not work with images. Can have problems with tables.
Do not work with dynamic web sites.