Systematic Comparison of Vectorization Methods in Classification Context

Krzeszewska, Urszula; Poniszewska-Marańda, Aneta; Ochelska-Mierzejewska, Joanna

doi:10.3390/app12105119

Cited by 16 publications

(8 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The count vectorizer is built unconditionally on the number of word events in the text. The count vectorization technique performs both the counting of how many times a token occurs and the process of tokenization [17]. The count vectorizer is based on various parameters that are used to purify the type of features.…”

Section: Feature Extraction 331 Count Vectorizermentioning

confidence: 99%

Artificial Intelligence Based Sentence Level Sentiment Analysis of COVID-19

Rukhsar¹,

Awan²,

Naseem³

et al. 2023

Computer Systems Science and Engineering

View full text Add to dashboard Cite

Web-blogging sites such as Twitter and Facebook are heavily influenced by emotions, sentiments, and data in the modern era. Twitter, a widely used microblogging site where individuals share their thoughts in the form of tweets, has become a major source for sentiment analysis. In recent years, there has been a significant increase in demand for sentiment analysis to identify and classify opinions or expressions in text or tweets. Opinions or expressions of people about a particular topic, situation, person, or product can be identified from sentences and divided into three categories: positive for good, negative for bad, and neutral for mixed or confusing opinions. The process of analyzing changes in sentiment and the combination of these categories is known as "sentiment analysis." In this study, sentiment analysis was performed on a dataset of 90,000 tweets using both deep learning and machine learning methods. The deep learning-based model long-short-term memory (LSTM) performed better than machine learning approaches. Long short-term memory achieved 87% accuracy, and the support vector machine (SVM) classifier achieved slightly worse results than LSTM at 86%. The study also tested binary classes of positive and negative, where LSTM and SVM both achieved 90% accuracy.

show abstract

Section: Feature Extraction 331 Count Vectorizermentioning

confidence: 99%

Artificial Intelligence Based Sentence Level Sentiment Analysis of COVID-19

Rukhsar¹,

Awan²,

Naseem³

et al. 2023

Computer Systems Science and Engineering

View full text Add to dashboard Cite

show abstract

“…Machine learning these days is one of the many technologies employed in medicine and particularly medical data analysis. To this day, there are countless approaches of machine learning that are applied in various medical problems, whether it is hospital readmission, diagnosis or treatment plans, and even newer methods and applications are being developed as of now [22][23][24][25].…”

Section: Related Work On Machine Learning For Medical Systemsmentioning

confidence: 99%

“…Feature extraction through NLP requires the application of multiple steps (Figure 3 [10,[16][17][18][19]22]):…”

Section: Figure 2 Transformation Of Text With Patient Notes To Table ...mentioning

confidence: 99%

Medical Data Transformations in Healthcare Systems with the Use of Natural Language Processing Algorithms

Poniszewska-Marańda

Vynogradnyk²,

Maranda³

2023

Applied Sciences

View full text Add to dashboard Cite

Machine learning has only recently begun to see its application in medicine and is still facing quite a few challenges that prevent it from being more widely used. Problems such as high data dimensionality and the lack of a common data schema still remain relevant. It is worth examining the usage of machine learning in the context of healthcare and deploying selected machine learning algorithms on the problem of cardiovascular disease diagnosis. Cardiovascular diseases are currently the most common cause of death in the world. Many of them develop for a long time in an asymptomatic way, and when the first symptoms become visible, it is often too late to implement effective treatment. For this reason, it is important to carry out regular diagnostic tests that will allow you to detect a given disease at an early stage. It is then possible to implement appropriate treatment that will prevent the occurrence of an advanced form of the disease. While doing so, it attempts to analyse data from different sources and utilizing natural language processing to combat data heterogeneity. The paper assesses the efficiency of various approaches of machine learning (i.e., TR-SVM (Terminated Ramp–Support Vector Machine), TWNFI (Transductive Neuro-Fuzzy Inference), Naive Bayes) when applied in the healthcare field and proposes the solutions to the problem of plain text data transformation and data heterogeneity with the help of natural language processing. The algorithms used for diagnosis were implemented, tested and their performance compared, with their parameters also investigated, making it easier to choose an algorithm better suited for a specific case. Whereas TRSVM is better suited for smaller datasets with a high amount of dimensions, TWNFI performs better on larger ones and does not have the performance problems.

show abstract

“…To do this, the key characteristics are assigned as elements of the vector. For example, if degree distribution, clustering coefficient, and betweenness centrality are chosen as key characteristics, then each network can be represented as a vector with these elements (Krzeszewska et al, 2022). The vectorization process is an important step because it enables the networks to be compared using mathematical techniques such as CS.…”

Section: Vectorizationmentioning

confidence: 99%

Analyzing complex networks: Extracting key characteristics and measuring structural similarities

Gul,

Al‐Obeidat,

Amin

et al. 2023

Expert Systems

View full text Add to dashboard Cite

SummaryThis paper discusses the importance of feature extraction and structure similarity measurement in the analysis of complex networks. Social networks, biological systems, and transportation networks are just a few examples of the many phenomena that have been modeled using complex networks. However, analyzing these networks can be challenging due to their large size and complexity. Feature extraction techniques can help to simplify the network by identifying key nodes or substructures. Structure similarity measurement techniques can be used to compare different networks and identify similarities and differences between them. Previous research has suggested that real‐world complex networks are influenced by multiplex features and either local or global features. However, the interaction between these characteristics is not well understood. The proposed approach outperforms other graph similarity methods on publicly available datasets, with accurate estimations of overall complex network structures. Specifically, the approach based on cosine similarity outperforms as compared to existing methods. Overall, this study highlights the importance of considering various graph features–local and global features and their interactions in the analysis of complex networks.

show abstract

Systematic Comparison of Vectorization Methods in Classification Context

Cited by 16 publications

References 14 publications

Artificial Intelligence Based Sentence Level Sentiment Analysis of COVID-19

Artificial Intelligence Based Sentence Level Sentiment Analysis of COVID-19

Medical Data Transformations in Healthcare Systems with the Use of Natural Language Processing Algorithms

Analyzing complex networks: Extracting key characteristics and measuring structural similarities

Contact Info

Product

Resources

About