Scribd
Upload
4 views3 pages

Optimizing Document Search Algorithms

The document discusses issues in existing document retrieval systems, highlighting the lack of optimization in search results. It introduces the Pegasos algorithm for optimizing SVM objective functions, demonstrating its effectiveness through empirical results. Additionally, it presents the Rank Semantic Association Algorithm to improve search result ranking based on user-defined parameters and relevance, addressing the limitations of current search methods.

Uploaded by

sruthi
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
4 views3 pages

Optimizing Document Search Algorithms

The document discusses issues in existing document retrieval systems, highlighting the lack of optimization in search results. It introduces the Pegasos algorithm for optimizing SVM objective functions, demonstrating its effectiveness through empirical results. Additionally, it presents the Rank Semantic Association Algorithm to improve search result ranking based on user-defined parameters and relevance, addressing the limitations of current search methods.

Uploaded by

sruthi
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Problems in Existing System:

Which searching the documents it retrieves the related contents. It doesn’t provide an
optimization result. For example we search about Tsunami it retrieves the all the details about
Tsunami, but we mention about particular task the query retrieves all the information but it
doesn’t matter about the time.

The Pegasos Algorithm

We describe the Pegasos algorithm for solving the optimization problem. The algorithm receives
as input two parameters: T - the number of iterations to perform; k - the number of examples to
use for calculating sub-gradients.

When the instances are sparse, namely, when each instance has very few non-zero elements. In
this case, we can represent w as a triplet (v, a, ν) where v is a dense vector and a, ν are scalars.
The vector w is defined through the triplet as follows: w = a v and ν stores the squared norm of
w, ν = kwk2. Using this representation, it is easily verified that the total number of operations
required for performing one iteration of Pegasos with k = 1 is O (d), where d is the number of
non-zero elements in x.

We described and analyzed a simple and effective algorithm for approximately minimizing the
objective function of SVM. The algorithm, called Pegasos, is a modified stochastic gradient
method in which every gradient descent step is accompanied with a projection step. We derived
fast rate of convergence results and experimented with the algorithm. Our empirical results
indicate that for linear kernels, Pegasos achieves state-of-the-art results, despite or because of its
simplicity. We plan to investigate all the questions we surfaced in this paper as well as to
conduct thorough experiments with non-linear kernels. In addition, we have started investigating
the usage of similar paradigms in other learning problems.

Rank Semantic Association Algorithm:

Problem is Existing System:

When searching the web it retrieves all the contents. It doesn’t provide an optimization result
based on the user Search.

Currently searched details only will displayed by using normal search

Ranking of the result not depending on the popularity of search content

To overcome this we use the following algorithm.


The complex relationships are based on property sequences that link the two entities in the
semantic association. Two entities ei and ej are semantically associated with each other. Next for
each of these entities we find the relevance, specificity and the user-specified span. The user
assigns weights for each of the parameters to refine the request. This also makes the ranking
process more flexible. Our current approach assigns binary values to the ranking parameters.
Assigning a range of specific values to these parameters is part of our future work

You might also like