http://dlib.iit.ac.lk/xmlui/handle/123456789/1350 2026-04-21T13:28:17Z http://dlib.iit.ac.lk/xmlui/handle/123456789/1416 Intelligent radio listenership monitoring platform Rodirgo, Suhash "The ability of the media to successfully shape public opinion is undeniable. Radio has always been widely considered as a part of the mainstream media in Sri Lanka. It has been recommended to have a complete radio listenership monitoring platform to monitor the listenership of each channel in light of the intense competition among the various radio channels in Sri Lanka. Currently, a diary-based system is used to monitor the radio listenership in Sri Lanka. However, the time has come to develop an advanced intelligent radio listenership monitoring platform to cater to this requirement which will enable better comprehension of the requirements and preferences of the general public. With the help of this research, machine learning technologies and other neural network technologies will be combined to create an intelligent radio listenership monitoring platform that is both reliable and accurate. This research will present a fool-proof approach of monitoring the radio listenership, which will resolve many of the practical challenges that have been a part of the current diary-based system. Radio channel heads advertisers and marketing businesses will find this important information supplied by the solution to be helpful in understanding the public interest and positioning their advertisements on appropriate segments. " 2022-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/1415 Detection of visitor expectations in the tourism industry using sentiment analysis during the COVID pandemic Amaratunge, Lahiru "The coronavirus disease 2019 (COVID-19) pandemic has had an unprecedented impact on the hotel industry, causing serious social and financial risks. The COVID-19 pandemic has had a negative impact on the hotel industry’s hospitality and challenged travel across the globe. The COVID-19 pandemic was between December 2019 and March 2021. Using sentiment analysis and topic modelling on client feedback regarding the hospitality offered by hotels during this time period in various countries, this investigation intends to identify customer satisfaction. Using sentiment analysis, it is possible to categorize customer satisfaction in a number of different ways. We developed improved parameters. Topic modelling is used to understand the various topics that customers discuss the most. We discovered that Indonesia and America are capable of meeting customer expectations. Sri Lanka performed well in Asia. We found that the top 14 topics that people were talking about were overall service, staff, cleanliness, room, slow booking, and hotel response to a pandemic. For topic modelling, we have selected simple LDA and LDA Mallet models. The LDA model has a coherence score of 0.35, whereas the LDA Mallet model has a coherence score of 0.49, which shows that the Mallet model separated the topics much better. So the LDA Mallet model is performing better than the normal LDA model. Senior hotel managers in developed and developing nations will benefit from the study's findings as they work to introduce innovative services that will please patrons and win back their trust." 2022-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/1414 Automatic content quality measurement of technical articles / blogs Perera, Andrea "The use of the internet is growing every day as technology advances. Technical articles/blogs are appealing to readers and researchers due to their ability to express a wide range of opinions and knowledge on a variety of topics and technology trends. As interest in how individuals obtain information changes, research on blog quality has grown in importance. The rapid expansion of this online environment creates a significant need for strengthening the quality of articles/blogs. In some instances, humans are still involved to assess the quality of article content and it’s a time-intensive process and requires more resources. Conventional methods only adopt page views or article popularity quality indexes to evaluate the quality of an article. Most experts examined how to improve the quality of the article based on SEO in order to place articles in the top search results, rather than the article's content. The author proposes a system that will focus on evaluating article quality on the article content-based features which are measuring article content breadth and depth along with other features which are the usage of valid URLs, Images, Tables/ diagrams, and code and usage of the Expertise/experience Personal opinions by giving a score. Content breadth score is a score/rating of how many related subjects/topics are covered within the article content and Content depth score is a score/rating of how detailed information coverage of a specific topic is within the article. Evaluated the proposed system using human annotated scores for content breadth and depth confusion matric along with the accuracy of 70% and 60%." 2022-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/1413 Classification and diagnosis of Covid - 19 and pneumonia using deep learning approach Epa, Sachini "Covid19 can be introduced as the factor that changed the lifestyle and the pattern of modern people, the most. Since its emergence in late 2019 in Wuhan, China, the Covid- 19 virus has spread all around the globe within a very short period resulting in millions of confirmed Covid-19 cases, increasing day by day. Although there are several ways of tests which can be performed to test for Covid-19 virus, most of the times, the amount of available test kits is not sufficient during the peaks of the covid-19 waves and the cost of the test kits is quite high to the extend where some people are not able to afford the cost. In addition, a medical professional’s help is needed to perform the existing tests, where the health professional is also at risk of getting exposed to an infection, while performing the tests. In addition to Covid-19, pneumonia can be identified as another major disease that occurs in the human respiratory system which often gets mistakenly identified, occurring at a considerably larger number of deaths a year. As per the aforementioned reasons, today, it is very much important to have a way that is able to identify both covid-19 and pneumonia which also addresses the issues that have arisen regarding the covid-19 tests, cost and the lack of required equipment. As a solution, alternative ways that can be used for the aforementioned purpose should be developed to address this issue and to minimise the social effects it has on the community. Therefore, this study attempts to introduce a new methodology to identify both covid- 19 and pneumonia given the chest x-ray of the patient which will identify the status of the given x-ray, stating whether the person is healthy or affected with covid-19 pneumonia, non-covid-19 pneumonia or bacterial pneumonia." 2022-01-01T00:00:00Z