This is a pre-production deployment of Warehouse, however changes made here WILL affect the production instance of PyPI.
Latest Version Dependencies status unknown Test status unknown Test coverage unknown
Project Description
# Near-Duplicate Detection

This program identifies near-duplicates in a corpus using techniques described
by Professor William Arms of Cornell University in his lectures to the students
of INFO 4300, Information Retrieval, Fall 2012.

This program was written by Parker Moore (pjm336), Fall 2012.

## Install

pip install git://

## Usage


## Explanation of Methodology

All of the logic for the program is built into the Detector class
(``). This class contains the methods and instance variables needed
to detect near-duplicates, such as the `get_jaccard(file1, file2)` method, the
`calculate_sketches()` method and the fundamental `create_3grams()` method.

This program implements the standard procedure for detecting near-duplicates:

1. Generate n-grams (3-grams in this case) for each document. Assign these
n-grams a unique ID based on a 64-bit hash.
2. Create 25 sketches for document based on 50 randomly selected
numbers and some stuff we generated earlier:
- `p` is the closest prime number to the # of n-grams
- `a_s` random, in the range [1, p-1]
- `b_s` random, in the range [0, p-1]
- `x` is the n-gram ID (the hash generated in step 1)
- using the equation: `f_s(x) = (a_s*x + b_s) % p`
- note: this equation is calculated 25 times per document (one time per
random pair `a_s` and `b_s`), but only the minimum result of
`f_s(x)` for each of the 25 pairs is saved. Thus, at the end of
the calculation, each document has 25 `f_min`'s, one for each
pair of random numbers.
3. Go through each document and compare it to all the other documents using the
Jaccard coefficient estimation equation : `J(d1, d2) = m/n`, where:
- `m` = number of sketch values (must be at the same index in respective
lists) which are the same between the two documents
- `n` = number of samples (# of sketches)
4. Having defined an arbitrary Jaccard coefficient threshold of `0.5`, the
program prints out the names of the documents whose Jaccard coefficient
is greater than the threshold previously defined, as well as the corresponding
Jaccard coefficient.

As an addendum to the project, the three "nearest neighbors" to the first ten
documents is calculated at the end using the same method (and the data from

## License

Standard MIT license applies.
Release History

Release History


This version

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

Download Files

Download Files

TODO: Brief introduction on what you do with files - including link to relevant help section.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
NearDuplicatesDetection-0.2.0.tar.gz (5.9 kB) Copy SHA256 Checksum SHA256 Source Mar 21, 2013

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting