Social media and online discussion forums have allowed people to hide behind a veil of anonymity, which has made the platforms feel unsafe for people with a different opinion than the majority. Recent research on robots and bots have found that they are a good option when it comes to inducing cooperation or acting as a conversation partner to encourage critical thinking. These robots and bots are based on an algorithm that is able to identify and classify comments, usually into positive and negative comments, left by users. The problem attended to in this thesis is to explore the possibility of creating an algorithm that can classify and pick out a minority opinion with an accuracy of at least 90%. The purpose is to create one of the vital algorithms for a larger project. The goal of this thesis is to provide a functioning algorithm with an accuracy of at least 90% for future implementations. In this thesis, the research approach is quantitative. The results show that it is possible to create an algorithm with the ability to classify and identify comments that also can pick out a minority opinion. Furthermore, the algorithm also achieved an accuracy of at least 90% when it comes to classification of comments, which makes the search for a minority opinion much easier. / Sociala medier och diskussionsforum online har tillåtit människor att gömma sig bakom sin datorskärm och vara anonym. Detta har gjort sociala medier till en osäker plats för människor som inte delar samma åsikt som majoriteten om olika diskussionsämnen. Ny forskning om robotar och sociala botar har funnit att dem är effektiva med att få människor att samarbeta samt att dem är en bra konversationspartner som framkallar mer kritiskt tänkande. Dessa robotar och sociala botar är baserade på en algoritm som kan identifiera och klassificera kommentarer, oftast till positiva eller negativa kommentarer som användare av sociala medier har lämnat. Problemet som avhandlingen försöker lösa är om det är möjligt att skapa en algoritm som kan identifiera och klassificera kommentarer, men även hitta och ta fram en åsikt som inte är en del av majoriteten med en träffsäkerhet på minst 90%. Ändamålet är att skapa en viktig byggsten för ett större forskningsprojekt. Målet med avhandlingen är att skapa en funktionerande algoritm för framtida undersökning som förhoppningsvis kan motarbeta partiskhet i sociala medier. Avhandlingens ståndpunkt är kvantitativ. Resultaten från avhandlingen visar att det är möjligt att skapa en algoritm som kan klassificera samt hitta en åsikt som inte är en del av majoriteten. Dessutom har algoritmen hög noggrannhet när det gäller klassificeringen vilket underlättar sökandet av en åsikt.

Background: In the last few years, there has been rapid growth in the use of cryptocurrency, as it is a form of digital currency and was developed using blockchain technology, so it is almost impossible to counterfeit cryptocurrency. Due to these features, it has attracted a lot of popularity and attention in the market. There has been a research gap in predicting accurate cryptocurrency prices by using sentiment analysis. This study will use Artificial Intelligence-based methods and sentiment analysis to develop a model for predicting cryptocurrency prices. By using the mentioned methods in this thesis, the developed model will provide precise results. Objectives: The objective of the thesis is to compare artificial intelligence models for cryptocurrency price prediction and analyze the importance of sentiment analysis by understanding the public pulse in cryptocurrencies and how it affects price fluctuations, analyzing the correlation within news articles, social media posts, and price fluctuations, as well as evaluating the model performance by employing metrics like RSME, MSE, MAE, and R2 error. Methods: The thesis follows the use of a systematic literature review along with an experimental model for comparing artificial intelligence models. Sentiment analysis played a crucial role in understanding market dynamics. By using linear regression, random forest, and gradient boosting algorithms artificial intelligence models are built to predict cryptocurrency prices using sentiment analysis. The developed models are then compared using performance metrics. This research has analyzed and evaluated each model's performance in predicting cryptocurrency prices. Results: The results of the systematic literature review indicated that market sentiment affects cryptocurrency prices. Prices have increased when market sentiment has been positive, whereas prices dropped when sentiment has been negative. The correlation between cryptocurrency values and market mood, however, is complicated as it depends on a variety of factors. Based on the evaluation measures, the random forest artificial intelligence model is the most accurate in predicting cryptocurrency prices after evaluating the three artificial intelligence models. Conclusions: This study utilized sentiment analysis and artificial intelligence to forecast cryptocurrency prices. It highlighted the significance of sentiment analysis as a tool for predicting the short-term price of cryptocurrencies by demonstrating how negative sentiment is correlated with decreases in price compared to positive sentiment with price increases. However, it recognized that it was necessary to take into consideration the complexity and broad range of effects on cryptocurrency markets. Research in the future will examine comprehensive sentiment analysis methods and broadening data sources.

No description available.

No description available.

No description available.

No description available.

Recently, distance education has become popular and has gotten much attention. Information and Communication Technology advances fostered distance learning creation and enabled individuals to participate in the education process via various web-based platforms and study entirely online. Thus, the notion of e-learning and distance learning emerged. Massive Open Online Courses (MOOCs) appeared as part of e-learning in 2008 and attracted great interest, especially during the COVID-19 pandemic. It was anticipated that this kind of study also could be integrated into higher education and revolutionize the learning approach. However, several issues related to MOOCs limit their full potential. One of the most significant problems is substantial rate of learners’ attrition. It was discovered that only 5-10 percent of MOOC learners complete a course. This thesis aims to examine what influences individuals’ decision to leave MOOCs and how learners perceive various course components to get ideas regarding how MOOCs could be enhanced. To do this, the mixed-method study was undertaken where quantitative data analysis of learners’ reviews from discussion forums and qualitative interviews were adopted. It allowed to get two perspectives and broaden the thesis out- come. For the current research, data was collected from six courses in three different subjects-«Health», «Art and Humanity/Design» and «Computer/Data Science». In the first part of the work, sentiment analysis and topic modeling using Python packages were carried out, and then the results were used to construct an interview questionnaire. Lexicon-based sentiment analysis technique and LDA topic modeling algorithm were utilized and proved to be robust methods to extract texts’ polarity and peoples’ opinions. In the qualitative part, 19 topics of discussion were identified, which were consolidated into eight topics with higher abstraction – materials, instructor, content, time, assignment, feedback, program(course), and algorithms. Then during the qualitative part, participants expressed their opinions regarding these topics, and analysis codes were predefined, and new topics did not emerge. The results showed learners’ perceptions related to presented topics and how these aspects influence experience with MOOCs. The outcome also showed a slight disparity between different subject learners, in both qualitative and quantitative studies identified topics of discussion were not exactly the same, showing that learners from different educational domains tend to discuss different themes.

In our work, we deploy semi-supervised learning methods to perform Sentiment Analysis on a corpus of sentences, meant to be labeled as either happy, neutral, sad, or angry. Sentence-BERT is used to obtain high-dimensional embeddings for the sentences in the training and testing sets, on which three classification methods are applied: the K-Nearest Neighbors classifier (KNN), Label Propagation, and Label Spreading. The latter two are graph-based classifying methods that are expected to provide better predictions compared to the supervised KNN, due to their ability to propagate labels of known data to similar (and spatially close) unknown data. In our study, we experiment with multiple combinations of labeled and unlabeled data, various hyperparameters, and 4 distinct classes of data, and we perform both binary and fine-grained classification tasks. A custom Radial Basis Function kernel is created for this study, in which Euclidean distance is replaced with Cosine Similarity, in order to correspond to the metric used in SentenceBERT. It is found that, for 2 out of 4 tasks, and more specifically 3-class and 2-class classification, the two graph-based algorithms outperform the chosen baseline, although the scores are not significantly higher. The supervised KNN classifier performs better for the second 3-class classification, as well as the 4-class classification, especially when using embeddings of lower dimensionality. The conclusions drawn from the results are, firstly, that the dataset used is most likely not quite suitable for graph creation, and, secondly, that larger volumes of labeled data should be used for further interpretation.

Throughout the COVID-19 pandemic, people have turned to social media to share their experiences with the coronavirus and their feelings regarding subjects like social distancing, mask-wearing, COVID-19 vaccines, and other related topics. The publicly available nature of these social media posts provides researchers the chance to obtain a consensus on an array of issues, topics, people, and entities. For the COVID-19 pandemic, this is valuable information that can prepare communities and governing bodies for future epidemics or events of a similar magnitude. However, clearly defining such a consensus can be difficult, especially if researchers want to limit the amount of bias they introduce. The process of sentiment analysis helps to address this need by categorizing text sources into one of three distinct polarities. Namely, those polarities are often positive, neutral, and negative. While sentiment analysis can take form as a completely manual task, this becomes incredibly burdensome for projects that involve substantial amounts of data. This thesis attempts to overcome this challenge by programmatically classifying the sentiment of COVID-19 posts from 10 social media and web-based forums using a multinomial Naive Bayes classifier. The unique and contrasting qualities of the social networks being analyzed provide a robust take on the public's perception of the pandemic that has not yet been offered up to the present.

No / The growing incidents of counterfeiting and associated economic and health consequences necessitate the development of active surveillance systems capable of producing timely and reliable information for all stake holders in the anti-counterfeiting fight. User generated content from social media platforms can provide early clues about product allergies, adverse events and product counterfeiting. This paper reports a work in progress with contributions including: the development of a framework for gathering and analyzing the views and experiences of users of drug and cosmetic products using machine learning, text mining and sentiment analysis; the application of the proposed framework on Facebook comments and data from Twitter for brand analysis, and the description of how to develop a product safety lexicon and training data for modeling a machine learning classifier for drug and cosmetic product sentiment prediction. The initial brand and product comparison results signify the usefulness of text mining and sentiment analysis on social media data while the use of machine learning classifier for predicting the sentiment orientation provides a useful tool for users, product manufacturers, regulatory and enforcement agencies to monitor brand or product sentiment trends in order to act in the event of sudden or significant rise in negative sentiment.