共计 4324 个字符,预计需要花费 11 分钟才能阅读完成。
COMP SCI 4094/4194/7094 – Distributed Databases and Data Mining
Assignment 1
DUE: 9pm Monday 29th April 2019
Important Notes
Handins:
– The deadline for submission of your assignment is 9pm Monday 29th April, 2019.
– You must do this assignment individually and make individual submissions.
– Your program should be coded in C++ and pass test runs on the two test files. The
sample input and output files are downloadable in“Assignments”of the course home
page (https://cs.adelaide.edu.au/users/honours/dddm/19s1-dddm-adelaide/).
– You need to use svn to upload and run your source code in the web submission system
following Web-submission instructions stated at the end of this sheet. You should
attach your name and student number in your submission.
– Late submissions will attract a penalty: the maximum mark you can obtain will be
reduced by 25% per day (or part thereof) past the due date or any extension you are
granted.
Marking scheme:
– 12 marks for testing on 4 randomly generated tests: 3 marks per test, where 1 mark
is for the affinity matrix AA, and 2 marks for the clustered affinity matrix CA.
– 3 marks for the structure of your code.
If you have any questions, please send them to the student discussion forum. This way you
can all help each other and everyone gets to see the answers.
The assignment
In this assignment you are required to implement the Bond Energy Algorithm of vertical fragmentation.
Your code should contains two separate procedures AA Generator and CA Generator,
where AA Generator takes the input of all attributes of a relation, a set of queries and their
access frequencies at different sites, and produces the output of an affinity matrix AA, and CA
Generator takes input of an affinity matrix AA and produces a clustered affinity matrix CA. For
description of the BEA algorithm, definitions of AA and CA, please see lecture slides/textbook.
In this assignment, the Attribute Affinity is measured by the extended Otsuka-Ochiai coef-
ficient (https://en.wikipedia.org/wiki/Yanosuke Otsuka) instead of the traditional method described
in the textbook. The following equations show the details of the computation, where n
is the number of attributes, and m is the number of sites, Aik is the number of times Attribute
Ai
is accessed by Query qk, considering of all sites. For the result of division, you must round
off the decimals to four decimal digits.
Example
For AA Generator:
Input
The relation, called PROJ, has the following features Ai:
Label Name
A1 PNO
A2 PNAME
A3 BUDGET
A4 LOC
Queries (qi):
q1: SELECT BUDGET FROM PROJ WHERE PNO=Value
q2: SELECT PNAME, BUDGET FROM PROJ
q3: SELECT PNAME FROM PROJ WHERE LOC=Value
q4: SELECT SUM(BUDGET) FROM PROJ WHERE LOC=Value
Access frequency matrix ACC, where Si denotes the i-th site:
S1 S2 S3
q1 15 20 10
q2 5 0 0
q3 25 25 25
q4 5 0 0
Output
The attribute affinity matrix AA:
A1 A2 A3 A4
A1 45 0 45 0
A2 0 80 5 75
A3 45 5 53 3
A4 0 75 3 78
For CA Generator:
Input
The attribute affinity matrix AA:
A1 A2 A3 A4
A1 45 0 45 0
A2 0 80 5 75
A3 45 5 53 3
A4 0 75 3 78
Output
The clustered affinity matrix CA:
A1 A3 A2 A4
A1 45 45 0 0
A3 45 53 5 3
A2 0 5 80 75
A4 0 3 75 78
Web-submission instructions
First, type the following command, all on one line (replacing xxxxxxx with your student
ID):svn mkdir – -parents -m “DDDM”
https://version-control.adelaide.edu.au/svn/axxxxxxx/2019/s1/dddm/assignment1
Then, check out this directory and add your files:
svn co https://version-control.adelaide.edu.au/svn/axxxxxxx/2019/s1/dddm/assignment1
cd assignment1
svn add AAGenerator.cpp
svn add CAGenerator.cpp
svn commit -m “assignment1 solution”
Next, go to the web submission system at:
https://cs.adelaide.edu.au/services/websubmission/
Navigate to 2019, Semester 1, Distributed Databases and Data Mining, Assignment 1.
Then, click Tab“Make Submission”for this assignment and indicate that you agree to the
declaration. The automark script will then check whether your code compiles. You can
make as many resubmissions as you like. If your final solution does not compile you will
not get any marks for this solution.
Note:i. The auto-marker script compiles and runs the two cpp files named“AAGenerator.cpp”
and“CAGenerator.cpp”one by one.
ii. The auto-marker script will compile your AAGenerator.cpp and CAGenerator.cpp by
the following command:
g++ -std=c++11 AAGenerator.cpp -o runAA g++ -std=c++11 CAGenerator.cpp -o
runCA
iii. Your AAGenerator.cpp should accept three input text files in the order of Attributes
(att), Queries (query) and Access Frequencies (acc), which are randomly generated
by the system, then output and print the required attribute affinity matrix (AA).
Your CAGenerator.cpp should accept an input affinity matrix AA provided by the
system rather than reading your AAGenerator’s output AA, then output and print
the clustered affinity matrix (CA) as the output. In this way of testing AA and CA
separately, your marks will be maximized — you will receive marks for your correct
CAGenerator coding even if your AAGenerator produces incorrect AA.
iv. The file path and the file name in your local machine will not work with our websubmission
system.
WX:codehelp