Multi-class E-mail Classification with a Semi-Supervised Approach Based on Automatic Feature Selection and Information Retrieval

Fernández, Juan Manuel; Errecalde, Marcelo Luis

doi:10.1007/978-3-031-14599-5_6

Cited by 1 publication

(1 citation statement)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In 2022 Fernández, J.M.et al [13] presented a multi-class E-mail classification based on feature selection and information retrieval. The primary features are chosen for each class from an initial data set of manually labeled emails using three techniques: TF-IDF, logistic regression, and SS3.…”

Section: Review On Email Classification Techniquesmentioning

confidence: 99%

Email Net: Efficient Email Classification Approach Based on Graph Similarity Measure

2023

IJIES

View full text Add to dashboard Cite

In this Internet Era, Email is a vital form of communication for many academic, personal, and professional users. Despite the availability of alternative means of communication, such as social networks, electronic messages, and mobile apps, email remains an integral part of communication. Email auto-management techniques are necessary for a variety of reasons, such as saving users valuable time, dealing with high-dimensional data, and making email communication easier and more accessible. In this work, a novel email net (improved elephant herd optimization and Graph similarity with the Jaccard index) technique has been proposed for efficient email classification based on graph similarity measure. Initially, the dataset is pre-processed using NLP techniques such as removing email signatures, removing punctuations, removing stop-words, lowercase conversion, tokenization, and stemming for removing irrelevant data. After pre-processing the feature are extracted using bag of words and term frequencyinverse document frequency (TF-IDF). These extracted features are given as input to improved Elephant herding optimization (EHO) for selecting the most relevant features to build a graph-based similarity index for classifying each category of e-mail. The proposed email net was tested on a benchmark dataset and a real-time dataset. Also, the proposed method's performance is compared with other classifiers. According to the experimental results, the proposed approach outperforms all other classifiers with a 98.82% of accuracy.

show abstract

Section: Review On Email Classification Techniquesmentioning

confidence: 99%

Email Net: Efficient Email Classification Approach Based on Graph Similarity Measure

2023

IJIES

View full text Add to dashboard Cite

show abstract

Multi-class E-mail Classification with a Semi-Supervised Approach Based on Automatic Feature Selection and Information Retrieval

Cited by 1 publication

References 18 publications

Email Net: Efficient Email Classification Approach Based on Graph Similarity Measure

Email Net: Efficient Email Classification Approach Based on Graph Similarity Measure

Contact Info

Product

Resources

About