Keyword Categorization using Statistical Methods

Abstract

Keyword analysis is a way to gain insight into market behaviour. It is a detailed analysis of words and phrases that are relevant to the selected area. Keyword analysis should be the first step in any search engine optimization, as it reveals what keywords users enter into search engines when searching the Internet. The keyword categorization process takes up almost half of the total analysis time, as it is not automated. There is currently no known tool in the online advertising market that facilitates keyword categorization. The main goal of this paper is to streamline the process of keyword analysis using selected statistical methods of machine learning applied in the categorization of a specific example.

Keyword decisions in sponsored search advertising: A literature review and research agenda

“…Although how to effectively organize keywords following search advertising structures (i.e., keyword assignment and grouping) is a critical operational-level issue, as far as we are aware, in the literature, less attention has been put to keyword assignment and grouping, except for recent works on keyword grouping (Li & Yang, 2020) and keyword categorization (Krasňanská et al, 2021;Polato et al, 2021).…”

“…Krasňanská et al (2021) applied a one-against-one LSVM-based method to classify keywords into multiple categories. The LSVM-based method can obtain a higher accuracy rate compared with other machine learning methods such as multinomial logistic regression and multinomial Naï ve Bayes.…”

“…Together with keyword performance parameters such as click-through rate, the SSA structure matters in keyword decisions(Li and Yang, 2020). In keyword categorization, statistical characteristics of keyword like TF-IDF has often been listed as input features(Krasňanská et al, 2021), and internal word structures are also used as the underlying framework for keyword representation(Polato et al, 2021).…”