Research Publications
Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/4194
This main community comprises five sub-communities, each representing the academic contribution made by SLIIT-affiliated personnel.
Browse
10 results
Filters
Advanced Search
Filter by
Settings
Search Results
Publication Embargo CricSquad: A System to Recommend Ideal Players to a Particular Match and Predict the Outcome of the Match(IEEE, 2023-06-12) Lekamge, E. L.; Wickramasinghe, K. R.; Gamage, S. E.; Thennakoon, T. M. K. L.; Haddela, P.S; Senaratne, SSelection of the cricket squad plays a very important role in the outcome of the match. This work is about selecting ideal players for a cricket match and predicting the outcome of the match according to the selected cricket team. A cricket squad consist of around 15 to 16 players, with different expertise in batting, bowling, fielding. To select players for the squad, points were calculated using a statistical approach considering player’s overall career data. And then for the further use of selecting players for the squad next match performance of each and every player were predicted using Machine Learning techniques. Association rule mining was used to find frequent winning player combinations with day/night, home/away, batting first/second, against different opponent combinations. Finally calculate points for each player in both teams, then predict the outcome of the match with classification algorithms by considering the calculated total points of each team and other factors such as toss outcome, batting inning, day night conditions and venue. As for the results, XG boost regressor has produced the highest R2 score of 0.92 for batsman runs prediction model while random forest regressor has produced the highest R2 score of 0.66 for bowler wickets prediction model. The Gradient Boost Classifier predicted the Outcome of a match with the highest accuracy of 0.92 while the K Nearest Neighbor achieved the lowest accuracy of 0.82 score.Publication Open Access Cryptocurrency Price Prediction: A Comparative Study using LSTM, GRU and Stacking Ensemble Algorithm for Time Series Forecasting(SLIIT, 2022-02-11) Ashikul Islam, M. DTechnology has significantly reshaped how humans interact with their tangible and intangible surroundings. Cryptocurrency is considered to be one of the most recent technological inventions which revolutionized how we perceive currencies and their functionality. It has become popular because of its safety, security and anonymity. However, volatility remains one of the major issues with cryptocurrencies to this day. Therefore, the primary aim of this paper is to develop LSTM (Long ShortTerm Memory), GRU (Gated Recurrent Units) and a Stacking Ensemble Learning algorithm that efficiently predicts the price of a cryptocurrency for a given period of time. The predictions are then observed and analysed to determine the comparative performance of the said algorithms.Publication Open Access Special Event Item Prediction System for Retails – Using Neural Network Approach.(SLIIT, 2022-02-11) Alwis, T; Pemarathna, PSelling and buying is the general process marketing field follows. Nowadays marketing field bonded with the modern technology, and it highly effected to field expandability. Marketing become fruitful when it achieves its key points which are called sales and profit. Mostly people are move to the retails because all the essentials and other things can buy from one place. There are many technological concepts involve with marketing field as an enhancement. Prediction processes, data analysis, item designing and profit calculation are some representatives for those concepts. This study is a prediction process, developed for retails using machine learning approaches. Item sales data analyzed and generated prediction results on set of items which are given maximum or expected profit margins and which items satisfy the customer most. Item suppliers are key stakeholder type a retail can have, there is a recommender system in this approach for suppliers and the recommendation is based on past sales data. There are certain types of machine learning approaches used in sales item prediction, sales item feature prediction, sales price prediction and etc. Novelty of this research is, it focused only special event items such as items in Christmas season, items specialized for Mother’s Day, Valentine Day, Sinhala, and Tamil new year and etc. This research process had completely followed the machine learning neural network concept. Recurrent Neural Network is subpart of neural networks and this research study followed up through this RNN method. Neural network had applied using a form of machine learning called deep learning. This model had worked on sequential data therefor LSTM (Long Short-Term Memory) layers were used and to avoid overfitting issue several dropout layers were used. The results prove neural network method has highest accuracy.Publication Embargo Mammogram-Based Cancer Detection Using Deep Convolutional Neural Network(IEEE, 2021-12-02) Thudawehewa, H; Rathnayake, P.; Thudawehewa, T; Silva, CBreast cancer is now a common health problem among most women. Breast cancer is the world’s second-largest cause of mortality for women, and it affects mostly women over the age of 50. The major reasons are that most women do not have proper knowledge about breast diseases/conditions, and the inability to detect abnormalities in the initial stages. A mammogram is one of the best imaging modalities recommended by doctors to diagnose breast cancers. Consultant radiologists are necessary for the identification of those breast pathologies by mammogram images. For a human, it takes some time to read and have an opinion about the condition. Also, the pandemic situation makes the diagnosis processes even more difficult due to the unavailability of doctors and other medical staff. Deep learning approaches are applied for breast cancer detection, and it helps radiologists to identify breast pathologies quickly and accurately. In this work, the mammogram images are collected using MIAS, DDSM, and INbreast databases. The proposed system identifies the location of the lump within the breast, if the lump is malignant or benign, the size of the lump, and the state of the nipple (It is abnormal or not). Convolutional Neural Network (CNN) method for classifying screening mammograms obtained outstanding performance compared to the previous methods. This CNN method produces 96.5% accuracy for breast tumor classification and produces the 80% accuracy for nipple classification.Publication Embargo Traffic Density Estimation and Traffic Control using Convolutional Neural Network(IEEE, 2019-12-05) Ikiriwatte, A. K; Perera, D. D. R; Samarakoon, S. M. M. C; Dissanayake, D. M. W. C. B; Rupasignhe, P. LThe existing traffic light control systems are inefficient due to the usage of predefined algorithms on offline data. This causes in numerous problems such as long delays and a wastage of energy. Estimation of traffic density indirectly affects in decreasing the high traffic congestion which will occur due to the less planning of transportation infrastructure and the policies. The goal of this research is to introduce an applicable method to improve the existing static traffic signal system into a dynamic system. As an approach we analyze the use of machine learning algorithms to measure the traffic density to tackle this research problem of high traffic congestion. The main target is to implement this system for the four-way junctions since it is a place where the possibility of having a traffic congestion seems to be high. With use of these traffic density estimation algorithm, crowd density estimation and signal handling we conduct experiments on minimizing the congestion at four-way junctions. We decided on using convolutional neural networks as an advanced machine learning method to increase the accuracy of the learning algorithm.Publication Embargo Mammogram-Based Cancer Detection Using Deep Convolutional Neural Network(IEEE, 2021-12-02) Thudawehewa, H; Silva, C; Rathnayake, p; Thudawehewa, TBreast cancer is now a common health problem among most women. Breast cancer is the world’s second-largest cause of mortality for women, and it affects mostly women over the age of 50. The major reasons are that most women do not have proper knowledge about breast diseases/conditions, and the inability to detect abnormalities in the initial stages. A mammogram is one of the best imaging modalities recommended by doctors to diagnose breast cancers. Consultant radiologists are necessary for the identification of those breast pathologies by mammogram images. For a human, it takes some time to read and have an opinion about the condition. Also, the pandemic situation makes the diagnosis processes even more difficult due to the unavailability of doctors and other medical staff. Deep learning approaches are applied for breast cancer detection, and it helps radiologists to identify breast pathologies quickly and accurately. In this work, the mammogram images are collected using MIAS, DDSM, and INbreast databases. The proposed system identifies the location of the lump within the breast, if the lump is malignant or benign, the size of the lump, and the state of the nipple (It is abnormal or not). Convolutional Neural Network (CNN) method for classifying screening mammograms obtained outstanding performance compared to the previous methods. This CNN method produces 96.5% accuracy for breast tumor classification and produces the 80% accuracy for nipple classification.Publication Embargo Real-Time Greenhouse Environmental Conditions Optimization Using Neural Network and Image Processing(IEEE, 2020-11-04) Wickramaarachchi, P; Balasooriya, N; Welipenne, L; Gunasekara, S; Jayakody, AAgricultural business is one of the biggest areas in world economy. With the growth of population losing agricultural lands is the major issue in world food production. Therefore, controlled environment agricultural systems under vertical farming have been introduced with greenhouses. Within greenhouses there is not a mechanism to continuously monitor the growing community and change the climate conditions. Existing systems only predict the required conditions for the plant and once predicted that value is provided to the plants continuously or change the values from season to season. To address these issues, a working prototype of an IoT based smart hydroponic system is introduced, which uses computer vision to gain maximum profits by growing a specific cultivation by providing endemic environmental conditions and addressing the problems over its growing process. There, this research presents a way of external environmental condition optimization. Regression type Feed Forward Neural Network is considered for this research to optimize the required conditions for tomato plants. Based on the current height of the plant, expected height for next 24 hours, and growth date of the plants neural networks predict the CO2, temperature and humidity level for next 24 hours with the accuracy of 88.33%, 89.21% and 92.65% respectively. The objectives of the research can be achieved by this retrieved results. The successful implementation of neural networks results a cost-effective modern farming solution for growers. This research will be supportive to attain a fundamental comprehension on the concept of the research area.Publication Embargo Ayurvedic Knowledge Sharing Platform with Sinhala Virtual Assistant(2019 1st International Conference on Advancements in Computing (ICAC), SLIIT, 2019-12-05) Jayalath, A.D.A.D.S.; Nadeeshan, P.V.D.; Amarawansh, T.G.A.G.D.; Jayasuriya, H.P.; Nawinna, D. P.Apart from western medicine methods Ayurveda medicinal system is a very huge and better resulting medicinal technique. In these Ayurveda methods identification of indigenous plants to predict the medicines is very important and must do very carefully. Generally main components that we use to identify a plant are leaf, flower, trunk and root etc. Among these features, we use images of leaves and flowers. To do this we are using deep learning based CNN approaches and machine learning and technologies. Those are OpenCV, and Tensorflow classification algorithm. According to the evidences that we gathered from surveys and interviews that we conducted with the responsible parties we could find out that lots of people don’t have much knowledge about indigenous medicinal plants and their Ayurveda treatment methods. To overcome this problem we implemented Ayurveda information centralized chatbot which is able to answer user’s questions relevant to the Ayurveda and indigenous medicinal plants. Chatbot will analyze the question that user asks and will provide answers according to that. Another useful feature of this system is it provides relevant information of Ayurveda doctors. So users can find doctors according to their needs and they are able to rate and give recommendations for the doctors. That will be help others to find doctors more easily and efficiently without any doubt.Publication Embargo A Geophone Based Surveillance System Using Neural Networks and IoT(2020 2nd International Conference on Advancements in Computing (ICAC), SLIIT, 2020-12-10) Supun Hettigoda, Chamath Jayaminda; Amarathunga, U.; Wijesundara, M.; Wijekoon, J.; Thaha, S.Securing our assets and properties from intruders and thieves has become increasingly challenging as intruders become technology aware. The most common approach to monitor physical assets is CCTV. However, this approach has a number of technical limitations in addition to the cost. The CCTV camera location is visible to the intruder and intruder can also identify possible blind spots in the CCTV coverage area. In this paper, we introduce a novel method to secure physical assets using Geophones, Neural Networks, and IoT Platforms. This can either be used stand alone or to complement existing CCTV systems. In this approach, the system monitors vibrations on ground to detect intruders. We have achieved up to 93.90% overall accuracy for person identification. The system is invisible to intruders and covers a large area with a smaller number of nodes, thereby reducing the cost of ownership.Publication Embargo A particle-swarm-optimized fuzzy-neural network for voice-controlled robot systems(IEEE, 2005-12-05) Chatterjee, A; Pulasinghe, K; Watanabe, K; Izumi, KThis paper shows the possible development of particle swarm optimization (PSO)-based fuzzy-neural networks (FNNs) that can be employed as an important building block in real robot systems, controlled by voice-based commands. The PSO is employed to train the FNNs that can accurately output the crisp control signals for the robot systems, based on fuzzy linguistic spoken language commands, issued by a user. The FNN is also trained to capture the user-spoken directive in the context of the present performance of the robot system. Hidden Markov model (HMM)-based automatic speech recognizers (ASRs) are developed, as part of the entire system, so that the system can identify important user directives from the running utterances. The system has been successfully employed in two real-life situations, namely: 1) for navigation of a mobile robot; and 2) for motion control of a redundant manipulator.