Opinions about the government’s response to forest fires have drawn many opinions from the community. One way for people to express their opinions is to use social media Twitter. This study conducted a sentiment analysis process on the government’s response to handling forest fires in Indonesia in 2019 with data sources from Twitter. The analysis was carried out on 6325 datasets written on Twitter on September 20, 2019, and then through the process of pre-processing, automating labeling and classification. The automate labeling process uses a Vader that automatically detects the negative or positive polarity of each data and then goes through the classification process using the KNN algorithm. The test results that were built using rapidminer tools showed an accuracy level of the KNN algorithm of 79.45%, the highest if compared to other classifier algorithms such as decision trees, naïve Bayes and random forests. The sentiment analysis process can almost run automatically without human touch because there is already automated labeling using Vader. Testing sentiment analysis related to the government’s response to forest fires can be analyzed using the KNN algorithm and lexicon polarity detection Vader can be done properly.

This study conducted a sentiment analysis of the impact of the Covid-19 pandemic in the economic sector on people's lives through social media Twitter. The analysis was carried out on 23777 tweet data collected from 13 states in Malaysia from 1 December 2019 to 17 June 2020. The research process went through 3 stages, namely pre-processing, labeling, and modeling. The pre-processing stage is collecting and cleaning data. Labeling in this study uses Vader sentiment polarity detection to provide an assessment of the sentiment of tweet data which is used as training data. The modeling stage means to test the sentiment data using the random forest algorithm plus the extraction count vectorizer and TF-IDF features as well as the N-gram selection feature. The test results show that the polarity of public sentiment in Malaysia is predominantly positive, which is 11,323 positive, 4105 neutral, and 8349 negative based on Vader labeling. The accuracy rate from the random forest modeling results was obtained 93.5 percent with TF-IDF and 1 gram.

Abstract. Purpose: Consumer opinion is one of the essential keys that affect the success of a product. Sentiment analysis of consumer opinion is needed to find out information about customer satisfaction for companies in the decision-making process. The traditional sentiment analysis process extracts a complete sentiment from a single sentence. However, it does not consist of only one sentiment in one sentence. The total number depends on the number of aspects that make up the sentence. Therefore, a sentiment analysis process is needed to pay attention to aspects.Methods: This research focuses on product reviews from Indonesian e-commerce on several aspects of sentiment. Uses fastText word embedding to avoid Out of Vocabulary in datasets and Gated Recurrent Units for aspect spread detection. Sentiment classification on aspects using the Memory Network method.Result: The experiment results showed that aspect-based sentiment classification predictions had an accuracy of 83% compared to 78% overall classification predictions for review texts, indicating that aspect-based sentiment analysis can improve model performance on product review classification predictions.Novelty: Most product reviews analysis use document-level classification to extract and predict sentiment reviews, aspect-based analysis can be applied to product reviews for better sentiment understanding, using Memory Network to store important information explicitly on aspects and polarity.

Computer vision research in detecting and classifying the subtype Acute Lymphoblastic Leukemia (ALL) has contributed to computer-aided diagnosis with improved accuracy. Another contribution is to serve as an assistant and second opinion for doctors and hematologists in diagnosing the ALL subtype.Early detection can also rely on computer-aided diagnosis to determine initial treatment. The purpose of this study is to review the progress of research in the detection and classification of ALL subtypes. The method's discussion focuses on the application of deep learning to the domain of object detection and classification. Motivations, challenges, and future research recommendations are thoroughly discussed to improve understanding and progress in this field of study. The study was carried out methodically by analyzing a collection of papers on the detection and classification of ALL subtypes published in science direct, IEEE, and PubMed from 2018 to 2022. The analysis of this paper field is included in the results of the selected paper. The paper selection from among 65 papers was based on inclusion and exclusion methods. Based on research methods and objectives, papers are divided into two large groups. The first group discusses the classification of ALL subtypes, while the second group discusses the detection of ALL subtypes. The discussion of prior research reveals some challenging issues and future work, such as the limited availability of the ALL subtypes dataset, the high computational complexity of the deep learning model, and further exploration of transformers in computer vision as a reference for research gaps that can contribute to future research.

Banks try to get profit from society in various ways. One way is to use long-term deposit investment offers. If the product offering process for potential investors is not carefully considered, it will waste resources. Therefore, this study analyzes the accuracy of the predictions of consumers who have a high chance of participating in this program. The dataset used is historical bank data provided by Kaggle. In previous research, accuracy prediction has been carried out, but the accuracy is still low because it does not use a method to balance the class. Better accuracy can be improved using LightGBM and SMOTE methods. The test results with the number of testing data as much as 6590 and training data as many as 32950 show the highest accuracy of 90.63%.