Survey : Study of the Improvement of IR and SEO for a Better E-Reputation

EasyChair Preprint 11160

Abstract

The exponential increase in the amount of information, its variety and its velocity requires addressing them with new, more powerful tools for accessing relevant information in a short time. As well as the birth of the concept of e-reputation on the Web requires the improvement of the content of sites for better referencing by serach engines. Our article is the state of the art on the research of classical techniques of artificial intelligence for improving semantic search such as latent semantic analysis (LSA), probabilistic latent semantic analysis (pLSA) or their improvement the Latent Dirichlet Allocation (LDA) and the new techniques such as deep learning or more recently Transformers (BERT, GPT, etc.) for improving information retrieval (IR) and search engine optimization (SEO) by integrating semantic aspect in order to improve the natural referencing of a company or for a better e-reputation of the company in the Web. So we study the techniques used in the most recent works on several aspects, such as the accuracy and relevance of the information, the performance and quality of the result and the speed of obtaining the information. And we are doing a comparative study on their benefits and limitations. In order to show which technique is the most effective. For guidance in our future work.

Keyphrases: IR, LDA, LSA, PLSA, SEO, deep learning, e-reputation, transformers

@booklet{EasyChair:11160,
  author    = {Djoudi Kaouthar and Alimazighi Zaia and Dellal-Hedjazi Badiâa},
  title     = {Survey : Study of the Improvement of IR and SEO for a Better E-Reputation},
  howpublished = {EasyChair Preprint 11160},
  year      = {EasyChair, 2023}}