Browsing by Author "Ekanayake, E.M.U.W.J.B."
Now showing 1 - 11 of 11
Results Per Page
Sort Options
Item Application of Image Processing and Neural Network Technique for Rice Grading(Uva Wellassa University of Sri Lanka, 2019-02) Wasana, P.L.D.; Fernando, P.S.R.; Lakmal, M.K.A.; Illoshini, P.A.A.; Wilson, R.S.I.; Ekanayake, E.M.U.W.J.B.When considering the agricultural industry, Rice is a principal food source in Asian countries. It is the most commonly and widely used grain in the local consumer market. Thus, analyzing the quality of rice is important. The quality of Rice will depend on the milling. Most mill owners do not have a proper method of measuring the quality of rice. Calculations are currently carried out using the Vernier Caliper in research centers, but it is a time-consuming task. Though there are some machines for automating this process, its usage is very low because of the high cost. Rice can primarily be classified based on colour and shape. Here we analyzed the two genres, Red Kekulu rice and Samba rice produced by Bombuwala rice research center. This paper introduces the rice classification method according to image processing approaches and neural network. Physical characteristics of the grain such as major and minor axis length, perimeter, area, colour and chalkiness are used to classify the rice grains. Identifying broken rice and wastages are the major objectives of the grading system. We have compared the proposed system results with manual measurements and visual observations. Matlab tool is used for image acquisition, preprocessing, segmentation, feature extraction and training the data set. This proposed grading system has scales such as premium, grade A, grade B and it was defined under supervision of rice researchers using broken rice and wastage content. The proposed image processing methods can reduce the time of operation and increase the accuracy. Finally, with the proposed grading system consumers will be able to receive information regarding the quality of the rice.Item Augmentative and Alternative Communication Application for Adults with Language Difficulties: An Application Developed in Sinhala Language(Uva Wellassa University of Sri Lanka, 2019-02) Gunawardana, D.A.Y.K.; Jayathunaga, R.M.; De Silva, A.H.H.G.; Ekanayake, E.M.U.W.J.B.; Wilson, R.L.S.Many adults can experience acquired disorders such as stroke, Parkinson, amyotrophic lateral sclerosis that can interfere with their ability to communicate with others. Currently, adult individuals who experience these kinds of difficulties need to rely on low- technology options such as printed out alphabet boards to express themselves. The number of words that a person used to communicate with others is much larger than the number of words printed on the boards, hence they face difficulties in communicating with the others. To address this issue, a text-based Augmentative & Alternative Communication system was developed in the Sinhala language for adult persons with speech disorders. The system comprises of three components keyboard layout for Sinhala fonts, next word prediction, and the text-to-speech converter. The methodology of the study was applied as follows. The requirements of the patients were collected from the Disability Rehabilitation Department at the Ragama Base Hospital. The requirements were analyzed case by case and a keyboard layout was designed by taking all the requirements into consideration. The most important module of this system is the next word predictor. This module assists a patient to predict the next word once he selects a word. The RNN neural network model was trained with a typical set of words that such a person often used. The word sequence was constructed from the requirements identified from the interviews with such patient and the health care professionals who have been working in this field for a significant period. A model was trained to perform text-tospeech using TensorFlow libraries. Once, the word predictor constructs the sentence, the whole sentence is converted into voice at once. The initial evaluation of the system was conducted only with patients who are being received treatments at the Ragama Hospital. The test results show that the system is able to communicate easily with patients in decent accuracy.Item Augmented Reality-Based Approach to Improve Learnability of Sri Lankan History(Uva Wellassa University of Sri Lanka, 2020) Muthuhewa, A.M.C.D.; Fernando, W.P.V.R.; Senanayake, S.H.D.; Ekanayake, E.M.U.W.J.B.; Ellepola, C.S.D.Augmented Reality (AR) being one of the most trending technologies in the world has offered many possibilities to improve existing traditional approaches in day to day tasks. In Sri Lanka, the existing model of education is mostly based on a passive learning system. But as the technology is growing, all the education systems are moving towards digital. The low-cost tablet PCs would replace textbooks soon. In Sri Lanka, there are some schools where the tablet PCs are already using as a learning element. However, there is not enough learning material to be used with these high-end devices and the curriculum of the local education is not optimized to take full advantage out of these devices. This study is focused on how AR-related applications can be used to improve the learnability of students, allow students to learn in an active learning environment in history, and archeological education using AR and evaluate the feasibility of implementing an active learning environment. In this study, the evaluation has been done using two main areas in archeological history education in Sri Lanka. Which are archeological places and arts and crafts that have an archeological value? The evaluation has been done using a mobile AR application-HistoriaAR. The researchers select a group of students to study a given lesson that covers the history and archeological aspects using textbooks and AR-based mobile application in two instances. After that, students were evaluated with a paper-based questionnaire and an in-build evaluation component in the mobile AR application, respectively. According to the results, the mean score of performance (7.13) using textbooks has been improved to mean score of (8.10) using AR related to archeological places. And mean score of performance (6.92) using textbooks has been improved to mean score of (7.63) using AR for to arts and crafts section. At the end of the study, results have proven that the performance of the learning of students can be improved using AR. Keywords: Augmented reality (AR), Sri Lankan education, History & Archeological studies, Active learItem Automated Essay Type Paper Marking System(Uva Wellassa University of Sri Lanka, 2020) Wedisa, M.A.R.; Siriwardhana, M.K.S.S.; Dayananda, P.G.C.N.; Pathirana, K.P.P.S.; Ekanayake, E.M.U.W.J.B.Automated paper marking is a very important research tool for the education evaluation process. Some researchers indicated that almost every study’s challenge was to get the semantic similarity of an essay rather than keyword matching. Another major problem is the lack of sufficient data that needed to train the system for a specific domain with a supervised learning approach and there are some issues with the unavailability of educator’s involvement with the scoring systems, also there were no studies that behave like a complete system. The automated scoring or evaluation for written student responses have been, and are still a highly interesting topic for natural language processing (NLP) and Machine Learning (ML) research. This study is focused on building a complete system that automates essay paper marking with a novel approach using NLP and ML. Primarily, researchers have used a hybrid approach to get the semantic similarity between two textual objects which contain word-vector-similarity, knowledge-based- similarity, and word-order-similarity. As one of the main advantages, our system uses an unsupervised learning approach, so that the system can work independently without training for a specific subject domain. The emerging of word embedding encouraged the calculation of the word-vector-similarity with Vector Space Model and cosine-similarity mechanisms. On the other hand, the word-net knowledge base was used to calculate the semantic distance between the documents and word-order-similarity played a major role in the accuracy of the final result. Also, machine learning techniques and a vast number of NLP techniques have been used for implementation. Besides, the proposed study contains an OCR to identify student's handwritten characters and also a website to easily interact with the system. In conclusion, the system was tested and evaluated with 30 samples of essays and the manual scores given by the educators. As a result, it indicated a strong positive correlation of (0.882) between manual scores and the system scores. Keywords: Automated essay scoring (AES), Natural language processing (NLP), Machine learning (ML), Optical character reader (OCR)Item A Chatbot for Online Investing and Earning Services for School and College Students(Uva Wellassa University of Sri Lanka, 2021) Daulatzai, A.; Kodithuwakku, K.A.V.M.; Ekanayake, E.M.U.W.J.B.In many countries education is not free, students need to pay in order to pursue their studies. Therefore,financial hardship can play negative impact on student‟s academic progress. However,they have numerous opportunities to avoid their financial stress through online investments and earnings, but normally students are not aware about them. Hence, this paper aims to present a chatbot that provides online earning and investment platforms by scanning student‟s profiles.Although students can browse google in order to know regarding internet based earning and investment tendencies, nevertheless , it‟s not only time consuming but searching in google in order to findout useful information that suitable to that specific user‟s need is a kind of art , hence, everyone is not an artist. Thus, to overcome such obstacles , we proposed a chatbot that have a strong natural language ability to interact with users , initially the chatbot asks set of questions in order know the specific student‟s profile in according to that the chatbot will recommend them online earning and investment platforms that is fitted to that particular student. In addition to that, the chatbot not only suggests online earning and investment platforms but it also educates students that how to use them in an effective manner. The main purpose of this research paper is to encourage students to defeat their financial issues and to motivate them to become future entrepreneurs. Moreover, to inspire students to invest in Sri lankan stock markets. On the other hand, there are few limitations, for example the stock prediction module only work for Sri Lankan context , the proposed chatbot has only bitcoin price prediction module and the chatbot has the ability to help students not in all online earning and investing opportunities but in many such as e-commerce, stock buying and selling , forex trading , freelancing , social media earning module, cryptocurrency, affiliate marketing, website flipping , dropshipping, dropservice, saas business module and blogging. In conclusion, this proposed system can help students to avoid time-consuming method of browsing google or youtube by using of this free of cost chatbot Keywords: Chatbot; Machine Learning; Online earning; Amazon Lex; FinanceItem Developing a Computer Software for Blending Black Tea(Uva Wellassa University of Sri Lanka, 2019) Basnayake, C.C.; Perera, G.A.A.R.; Ekanayake, E.M.U.W.J.B.Tea is manufactured from the young shoot of the tea plant (Camellia sinensis L.). Black tea is manufactured by subjecting the fresh tea leaf in to withering, rolling, fermenting and drying. Quality of black tea varies with variety of tea, manufacturing process, region, climatic conditions etc. Price of black tea varies with its quality. Buyers of the bulk tea blend tea from various manufacturers to make the quality and price of the tea in the retail market fairly consistent. Quantity of tea from various manufacturers that should be added to make a particular blend is decided arbitrary by tea tasters based on the sensory evaluation of samples of tea from individual manufactures. Initially, a trial blend is prepared and check whether the intended properties are met. This process is time consuming and laborious. Therefore, user friendly computer software was developed using JAVA language to make the tea blending process convenient and accurate. Programing was done using matrix algorithm. Price and rank data (five point hedonic) on sensory evaluation (color brightness, strength and aroma) of initial tea samples and desired price and rank data (five point hedonic) on desired sensory properties of the intended blend are the input variables. The developed software is to be tested in commercial blending process and further fine-tuned.Item Landmark Recognition using Image Processing and Machine Learning(Uva Wellassa University of Sri Lanka, 2019-02) Fernando, K.P.D; Perera, M.I.U; Lakshmen, K.T.S.R; Ekanayake, E.M.U.W.J.B.In the modern world, tourism has become one of the fastest growing industries. In every form of tourism, the tourists encounter landmarks, which they have no knowledge on them. The current way of identifying the landmarks is either to refer printed material such as books, magazines, which describe the important landmarks of the particular region, refer the Internet, or get the assistance from a tour guide. Referring printed material while traveling is not a practical solution in today’s world. Referring the Internet may be practical, but after the landmark is accurately identified. Getting the assistance of a native person will involve financial costs and the source may be less reliable. As a solution to the above problems, we suggest a mobile application, which identifies landmarks using pictures and gives all relevant information, which would be useful for a tourist. The proposed solution uses image processing and machine learning to identify landmarks. A dataset of 5000 different landscapes is used in this project. The dataset was preprocessed using Caffe deep learning framework in order to remove unnecessary noise. One third of the dataset is randomly selected as the test set. The dataset was divided according to regions and models were trained for each region. SVM (Support vector machine), BOW (Bag of Words) and CNN (Convolutional Neural Network) algorithms were trained. The accuracy of the CNN model is 90%, while the accuracies of SVM and BoW are 70% and 60% respectively. Hence, the CNN model is used in this project. Dataset was divided according to regions and models were obtained for each region. Dividing the dataset into regions increased the accuracy and reduced the training time. The GPS data is used to identify the region and the appropriate model is used to retrieve the related information for the given landscape. The model was tested with real users and their positive feedback indicates the success of this project.Item Mobile Apps’ Feature Extraction Based On User Reviews Using Machine Learning(Uva Wellassa University of Sri Lanka, 2019-02) Thiviya, T.; Nitheesram, R.; Srinath, G.; Ekanayake, E.M.U.W.J.B.; Mehendran, Y.The star rating and user reviews of Google Play store play a major role in App Store Optimization. The average number of stars received for an app and the user reviews are used to evaluate the overall app quality. We argue that the star rating is not a reliable measurement for the user satisfaction, since the star rating is a straightforward mathematical expression only. We cannot find a user’s real experience about the app by asking them to rate the app out of five stars, even rating user reviews will give only an overall user perspective about the app. Therefore, we recommend a specific app feature evaluation method based on user reviews, which give us a genuine app rating than the conventional method. In addition to review based rating we did popular feature extraction from user reviews. Initially we started our research with mining user reviews from Google Play Store for several categories by using Web Scraping tool. We used the sentiment analysis to extract the meaning from the reviews and define the polarity of them. According to the polarity strength we rated each reviews. Overall rating was calculated by finding the average of given review based ratings. Then we compared review based ratings with the existing star ratings. We found that compare to overall review ratings, the existing star ratings differs and high. Moreover, the app feature set was selected according to the category of the app, and then the popularity of those features was calculated by using machine learning. We were resulted with app features and their popularity based on users' reviews. These popularities can be helpful to the potential app users to get know the top features and their popularity of the particular apps. In the meantime, app developers can identify which features have low popularity and they can improve those features in future.Item Movie Success and Rating Prediction Using Data Mining Algorithms(Uva Wellassa University of Sri Lanka, 2020) Pirunthavi, S.; Vithusia, R.P.; Abishankar, K.; Ekanayake, E.M.U.W.J.B.; Yanusha, M.This project developed the models to predict the success and the ratings of a new movie before its release. Since the success of a movie is highly influenced by the actor, actress, director, music director, and production company, those historical data were extracted from the Internet Movie Database (IMDb). The Box Office Mojo stores information about the cost of production of a movie and the total income of the movie. This information is helpful to determine whether the movie is successful or not in terms of revenue. A threshold was defined on revenue based on heuristics to categorize the movie into success or failure. Teasers’ and trailers’ comments were extracted from YouTube as those are very helpful to rate a movie. The keywords were extracted from the user reviews using a Natural Language Processing (NLP) technique and those reviews were categorized into positive or negative based on the sentimental analysis. A Random Forest Algorithm was trained using the features extracted from IMDb to predict the success of a movie. Further, the Naïve Bayers model was trained using the user reviews extracted from YouTube to predict the rating of a movie. The models were tested on real datasets and the accuracy of those were evaluated respectively. Finally, two conclusions have been met that the rating of a new movie cannot be predicted in advance through the YouTube trailers’ and teasers’ comments and the success of a new movie can be predicted in advance by using the data or features collected from online. The performances of the models are decent enough compared to the existing models in the literature. The Success Prediction model can be used as an early assessment tool of movies since it has gained 70% overall accuracy and hence, useful for the people in the movie industry and the audience of the movies. YouTube allows us to extract a limited number of user comments and hence, this factor could be negatively affected by the accuracy of the movie rating prediction. Keywords: Rating prediction, Data mining algorithmsItem A Nutrient Based Diet Plan Recommendation System using Machine Learning(Uva Wellassa University of Sri Lanka, 2021) Baskaran, K.; Yokarasa, K.; Paraloganathan, V.; Ekanayake, E.M.U.W.J.B.; Wilson, R.S.I.At present obesity is a key health issue as everyone is busy with their day-to-day lives. Existing diet recommendation systems suggest a common diet plan instead of considering the person‟s lifestyle and diseases and hence it leads to health issues. This research develops a system to recommend an appropriate diet plan for each person based on their personal profiles. The proposed system collects the personal information from users such as age, height, weight, gender, chronic diseases, and physical activities, and then it recommends the diet plans for the breakfast, lunch, tea time and dinner with appropriate calorie levels (carbohydrate, protein, lipid, calcium, phosphorous, fiber and iron) that helps to maintain the healthy weight of the body. The data was collected from the hospitals using a questionnaire. A Linear Regression models and a Neural Network model are trained to predict the required amount of calories per day based on the users‟ profile. Based on the error rate comparison of both model, the Neural Network model is the best fit for calorie prediction. The diet plan is defined by a rule-based system based on the predicted calorie level. The predicted diet plan for a given user is compared with the diet plan recommended by a nutritionist to measure the accuracy of the proposed system. Accordingly, the prediction accuracy of the system is 95%, which is decent enough when compared to the existing models in the literature. A limited number of parameters of users are considered to predict the calorie level and the diet food combinations. However, considering more parameters would further enhance the diet plan suggestions. Keywords: Machine Learning; Obesity; Linear Regression; Neural Network; Diet PlanItem Potentials to Develop Artificial Intelligence for the Hotel Industry in Sri Lanka (Special Reference to Star Class Hotels in Colombo District)(Uva Wellassa University of Sri Lanka, 2021) Dananjaya, B.W.D.; Ranasinghe, J.P.R.C.; Ekanayake, E.M.U.W.J.B.Artificial Intelligence (AI) has made great changes in various industries in response to the incessant technological revolution. Equally, such changes have affected the global hospitality and tourism industry at large. AI technology was originally used in the aviation industry and subsequently in the hospitality and tourism industry. It was used not only to satisfy the customers‟ needs but also as a mechanism to reduce cost and increase revenue. Furthermore, AI technology applied in the hospitality and tourism industry as it was used in the aviation industry. Currently, this technology is used globally in the hotel industry. Highly competitive and unpredictable challenges in the hotel operations have created the need of applying new technology to provide better services for their customers and maximize the efficiency and effectiveness of the Sri Lankan hotel industry too. The key objectives of this study were to examine different AI tools applied in the hotels in Sri Lanka, examining the perception of managers in the hotels on AI and identifying the barriers for AI implementation in hotels in Sri Lanka. A sample of ten General Managers and IT Managers from star graded hotels from Colombo district was drawn for the study using purposive sampling technique. Structured interview method was used to investigate perceptions, potentiality of implementation as well as barriers of implementation. Primary data was transcribed and analyzed using thematic analysis. Results indicate that, Sri Lankan hotels are not much familiar with the AI technology and have limited knowledge on AI application. All participants had a positive perception on AI technology use in the hotel industry. The paybacks of AI technology, rewards of AI to the hotel operations and facilitating factors that are needed for hotels, also were identified in this study. Based on the findings, researchers recommend fully developed AI technology for new and up-coming hotels for improved performance. AI tools can also be used in the existing hotels in Sri Lanka. Keywords: Artificial Intelligence, Benefits and Barriers, Hotel Industry, AI Tools, Maximize Efficiency and Effectiveness, Provide Better Service