Impact of corpus domain for sentiment classification: An evaluation study using supervised machine learning techniques

Karsi, Redouane; Zaim, Mounia; Alami, Jamila El

doi:10.1088/1742-6596/870/1/012005

Cited by 4 publications

(2 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The SVM method is applied to all types of features used in our experience with 90% of the dataset goes to the training set, and the remaining 10% to the testing set. SVM is an ideal choice, as it performs well on different domains [32].…”

Section: Resultsmentioning

confidence: 99%

Early Detection of Severe Flu Outbreaks using Contextual Word Embeddings

Karsi¹,

Zaim²,

Alami³

2021

IJACSA

Self Cite

View full text Add to dashboard Cite

The purpose of automated health surveillance systems is to predict the emergence of a disease. In most cases, these systems use a text categorization model to classify any clinical text into a category corresponding to an illness. The problem arises when the target classes refer to diseases sharing multiple information such as symptoms. Thus, the classifier will have difficulty discriminating the disease under surveillance from other conditions of the same family, causing an increase in misclassification rate. Clinical texts contain keywords carrying relevant information to distinguish diseases with similar symptoms. However, these specific words are rare and sparse. Therefore, they have a minor impact on machine learning models' performance. Assuming that emphasizing specific terms contributes to improving classification performance, we propose an algorithm that enriches training samples with terms semantically similar to specific terms using the deep contextualized word embeddings ELMo. Next, we devise a weighting scheme combining chi-square and semantic scores to reflect the relatedness between features and the disease under surveillance. We evaluate our model using the SVM algorithm trained on i2b2 dataset supplemented by documents collected from Ibn Sina hospital in Rabat. Experimental results show a clear improvement in classification performance than baseline methods with an F-measure reaching 86.54%.

show abstract

Section: Resultsmentioning

confidence: 99%

Early Detection of Severe Flu Outbreaks using Contextual Word Embeddings

Karsi¹,

Zaim²,

Alami³

2021

IJACSA

Self Cite

View full text Add to dashboard Cite

show abstract

“…Support vector machines: Support vector machines (SVM) is a supervised learning algorithm used to perform classification and regression tasks [25]. SVM is a classifier based on a statistical approach for either learning or prediction.…”

mentioning

confidence: 99%

Assessing naive Bayes and support vector machine performance in sentiment classification on a big data platform

Karsi

Zaim

Alami

2021

IJ-AI

Self Cite

View full text Add to dashboard Cite

Nowadays, mining user reviews becomes a very useful mean for decision making in several areas. Traditionally, machine learning algorithms have been widely and effectively used to analyze user’s opinions on a limited volume of data. In the case of massive data, powerful hardware resources (CPU, memory, and storage) are essential for dealing with the whole data processing phases including, collection, pre-processing, and learning in an optimal time. Several big data technologies have emerged to efficiently process massive data, like Apache Spark, which is a distributed framework for data processing that provides libraries implementing several machine learning algorithms. In order to evaluate the performance of Apache Spark's machine learning library (MLlib) on a large volume of data, classification accuracies and processing time of two machine learning algorithms implemented in spark: naive Bayes and support vector machine (SVM) are compared to the performance achieved by the standard implementation of these two algorithms on large different size datasets built from movie reviews. The results of our experiment show that the performance of classifiers running under spark is higher than traditional ones and reaches F-measure greater than 84%. At the same time, we found that under spark framework, the learning time is relatively low.

show abstract