A comparison of text classification methods using different stemming techniques

Bounabi, Mariem; Moutaouakil, Karim El; Satori, Khalid

doi:10.1504/ijcat.2019.101171

Cited by 19 publications

(15 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There exist several measures to evaluate the performance of a given classifier. Table 1 summarizes the set of most used evaluation measures (Bounabi et al , 2019).…”

Section: Experimentation and Resultsmentioning

confidence: 99%

“…Several processes govern the text classification operation (Bounabi et al , 2019), and the term weighting process has an essential impact on the classification quality, where a suitable weighting method gives the best classification accuracy. Using the NTF-IDF weighting term, we produce the vectors descriptors for o set of data set in natural language.…”

Section: Machine Learning Tools For Text Classificationmentioning

confidence: 99%

“…Generally, the representation and selection of features (Rajman and Besançon, 1998) affect the performance of the text mining tasks (Ge and Moh, 2017), e.g. a valuable vectorization of texts has a positive impact on the text classification assignment (Liu et al , 2018; Bounabi et al , 2019). One of the popular weighting methods is the term frequency–inverse term frequency (TF-IDF), and the FIS was used to calculate fuzzy TF-IDF (FTF-IDF), where the fuzzy sets and membership degrees are applied to define the document terms' fuzzy weights (Gupta et al , 2015).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A new neutrosophic TF-IDF term weighting for text mining tasks: text classification use case

Bounabi

Elmoutaouakil

Satori

2021

IJWIS

Self Cite

View full text Add to dashboard Cite

Purpose This paper aims to present a new term weighting approach for text classification as a text mining task. The original method, neutrosophic term frequency – inverse term frequency (NTF-IDF), is an extended version of the popular fuzzy TF-IDF (FTF-IDF) and uses the neutrosophic reasoning to analyze and generate weights for terms in natural languages. The paper also propose a comparative study between the popular FTF-IDF and NTF-IDF and their impacts on different machine learning (ML) classifiers for document categorization goals. Design/methodology/approach After preprocessing textual data, the original Neutrosophic TF-IDF applies the neutrosophic inference system (NIS) to produce weights for terms representing a document. Using the local frequency TF, global frequency IDF and text N's length as NIS inputs, this study generate two neutrosophic weights for a given term. The first measure provides information on the relevance degree for a word, and the second one represents their ambiguity degree. Next, the Zhang combination function is applied to combine neutrosophic weights outputs and present the final term weight, inserted in the document's representative vector. To analyze the NTF-IDF impact on the classification phase, this study uses a set of ML algorithms. Findings Practicing the neutrosophic logic (NL) characteristics, the authors have been able to study the ambiguity of the terms and their degree of relevance to represent a document. NL's choice has proven its effectiveness in defining significant text vectorization weights, especially for text classification tasks. The experimentation part demonstrates that the new method positively impacts the categorization. Moreover, the adopted system's recognition rate is higher than 91%, an accuracy score not attained using the FTF-IDF. Also, using benchmarked data sets, in different text mining fields, and many ML classifiers, i.e. SVM and Feed-Forward Network, and applying the proposed term scores NTF-IDF improves the accuracy by 10%. Originality/value The novelty of this paper lies in two aspects. First, a new term weighting method, which uses the term frequencies as components to define the relevance and the ambiguity of term; second, the application of NL to infer weights is considered as an original model in this paper, which also aims to correct the shortcomings of the FTF-IDF which uses fuzzy logic and its drawbacks. The introduced technique was combined with different ML models to improve the accuracy and relevance of the obtained feature vectors to fed the classification mechanism.

show abstract

“…There exist several measures to evaluate the performance of a given classifier. Table 1 summarizes the set of most used evaluation measures (Bounabi et al , 2019).…”

Section: Experimentation and Resultsmentioning

confidence: 99%

Section: Machine Learning Tools For Text Classificationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A new neutrosophic TF-IDF term weighting for text mining tasks: text classification use case

Bounabi

Elmoutaouakil

Satori

2021

IJWIS

Self Cite

View full text Add to dashboard Cite

show abstract

“…The stop words were eliminated using a stop words list. Finally, the Stemming algorithms are required to find the radicals of terms [3], where we use the Lovin stemming algorithm [3]. Globally, the text pretreatment helps the next used models to predict the most results using relevant data.…”

Section: Pretreatmentmentioning

confidence: 99%

“…A set of process governs the classification systems efficiently, such as the preprocessing of the used corpus [3]and the matrix embedding or the term weighting [4]. The first process permits a generation of the based vocabulary to produce one of the existing representation of descriptors, i.e., Boolean type [5], vector type [6], or probabilistic type [7].…”

Section: Introductionmentioning

confidence: 99%

The Impact of Neural Embedding Characteristics on Text Mining Tasks: Document Classification Use Case

Bounabi¹

2020

IJATCSE

Self Cite

View full text Add to dashboard Cite

One of the relevant text mining tasks is the document classification, where a useful content categorization control in many domains like content analyses, retrieval information, and the recommendation systems. In general, a set of process influence the classification system effectiveness, and the data representation has an essential impact on the text categorization as we will discover in this article. Hence, the paper's goal is to adjust the Paragraph Vector-Distributed Memory (PV-DM) as a variant of the current methods for neural text representation by comparing diverse neural parameters choices control the system complexity, e.g., epoch number, and vector size. Also, we employ a collection of classifiers subsequently combined using majority voting to show the impact of the neural PV-DM embedding on the binary business sentiment analysis, and multi labeled News data classification. The experiments prove that a suitable selection of the neural embedding characteristics enhances the hybrid machine learning model to 99% accuracy for a data type.

show abstract