A Comparative Study of Artificial Neural Network and Genetic Algorithm in Search Engine Optimization

Abstract

Search engine optimization applies search principles in search engines to assign a higher ranking to the most suitable webpage.  Nowadays, information searching is done ubiquitously on the World Wide Web with the help of search engines. However, the process needs to be efficient and produces accurate results at the same time. In this research, the objectives are to implement and evaluate the Artificial Neural Network and Genetic Algorithms. The accuracy result for both algorithms is compared by implementing keyword ranking, Search Engine Result Page visibility and time retrieval for document-based and e-commerce websites. To achieve them, firstly the problem and data are defined. Next, two datasets are imported from Kaggle and transformed into a more useful format. Then, the Artificial Neural Network and Genetic Algorithms are implemented on these datasets in Python using Jupyter Notebook tools. Subsequently, the accuracy of keyword ranking, Search Engine Result Page visibility and time retrieval for these datasets are observed based on the output and graph displayed. Lastly, an analysis of the results is performed. Conclusively, the Genetic Algorithm demonstrates a higher percentage of accuracy results than Artificial Neural Network algorithm in keyword ranking and SERP visibility. However, the accuracy results of time retrieval are vice versa. The results in Genetic Algorithm shows 9.0%, 9.0% and 3.0% in e-commerce dataset for keyword ranking and 4.0%, 51.0% and 1.0% in document-based dataset for SERP visibility. Next, Artificial Neural Network algorithm shows result 8.0%, 7.0% and 7.0% in e-commerce dataset and 3.0%, 50.0% and 4.0% in document-based dataset for time retrieval. Therefore, the results validated the ability of the Genetic Algorithm as one of the most applied algorithms in the search engine optimization field.