ICACT 2018

Permanent URI for this collectionhttp://repository.kln.ac.lk/handle/123456789/18944

Browse

Now showing 1 - 20 of 35

Age and Gender Related Variations in Human EEG Signals
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Pradeep, H.B.A.C.; Meegama, R.G.N.
There is a significant amount of research focused on discovering the functional behavior inside the human brain and methods to collect brain waves with respect to age. However, due to the lack of research using data-mining and pattern extraction methodologies on such data streams, we may be losing important features from human brain wave pattern data. The proposed research is aimed at collecting different kinds of brain wave patterns from different age categories of human beings and analyzing the correlation between the wave patterns of individuals. All the EEG data were taken from publically available and trusted data sources. The data from 22 subjects, five males and 17 females, within the age range from 3 to 22 years and were recorded with 256Hz and 16-bit resolution. We used FP1 and F7 channels as our main data sources for comparing and classification purposes. In the first phase, we applied a filtering process to clean the EEG data set of young male and female subjects to extract the hidden patterns. As EEG signals are acquired as a continuous stream, we use the sliding dot product or sliding inner product of two wave forms while searching for a long signal for shorter, known feature which is referred to as cross correlation. A correlation function is a function that gives the statistical correlation between random variables. In our research, the correlation between two signal forms (data sets) was used to measure the similarity between two wave forms. Subsequently, the cross correlation between all data pairs was calculated to find hidden relationships between each data group. In the sampling process, We ignored the first 256 data samples which was captured during 1s - 2s time period to compensate for possible errors added to the main brain wave during head movements and early adjustments. Using cross correlation diagrams, we observed similarity of brain wave signals between 11 year male and 22 year female subjects having a peak value of 3.5597e.
Air Pollution Monitoring System Using Arduino
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Rishan, U.M.
Arduino based Air pollution system is presented. Air pollution monitoring is old but very useful concept in day to day life. The level of air pollution has increased with times by lot of factors like the increase in population, industrialization, increased vehicle use and urbanization. Air pollution will directly affecting health of population. However the fresh air is necessary for all human being. Actually air pollution monitoring started from early using traditional way but the most sophisticated computer has been used to monitor the air quality. However in this project I am going to make an IOT based air pollution system using Arduino this will monitor the air quality accurately. The main objectives of this project to develop low-cost and ubiquitous sensor networks to collect real time data of urban environment. This air pollution system is connected with internet and we can monitor the air quality over the web server using internet. The alarm also embedded with this system that will trigger when the air quality goes down beyond a certain level, this means there are sufficient amount of harmful gases are present in the air like CO2, alcohol, and NH3.It will display the air quality in PPM on the LCD display and as well as on webpage so that we can monitor it very easily. In this IOT project, you can monitor the pollution level from anywhere using your computer or mobile devices.
Altered Brain Wiring in Alzheimer’s: A Structural Network Analysis using Diffusion MR Imaging
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Mahadevan, J.; Ratnarajah, N.; Ranaweera, R.D.
Alzheimer’s disease is a chronic neurodegenerative disorder and the most common form of dementia. It is characterized by cortical atrophy and disrupted anatomical connectivity as white matter fibre tracts lose axons and myelin degenerates. Biomarker tests are crucial to identify the early stages of the disease. It is currently a key priority in Alzheimer’s research to develop neuroimaging biomarkers that can accurately identify individuals in any clinical stage of the disease. Magnetic resonance imaging (MRI) can be considered the preferred neuroimaging examination for Alzheimer’s disease because it allows for accurate measurement of the 3-dimensional volume of brain structures. Diffusion Magnetic Resonance Imaging (DMRI), one of the methods, provides insights into aspects of brain anatomy that could never previously be studied in living humans. A comprehensive study of structural brain network in Alzheimer’s has been developed using diffusion MR imaging and graph theory algorithms, that can assess the white matter connections within the brain, revealing how neural pathways damaged in Alzheimer’s disease. A range of measurements of the network properties were calculated and the pattern of the community structure and the hub regions of the network were inspected. Global measures of efficiencies, clustering coefficients and characteristic path length confirms the disrupted overall brain network connectivity of Alzheimer’s. Relatively the same pattern of hub regions is preserved in Alzheimer’s, however, non-hub regions are affected, which indicates disease alters the internal pattern of the network especially the community structure. Modular analysis confirms this alteration and produces a different modular structure and increased number of modules in Alzheimer’s. Regional connectivity measures also indicated this change and the measures demonstrated the network centrality shifted from right hemisphere to left in Alzheimer’s. The knowledge gained from this study will support to find the strong imaging biomarkers of the Alzheimer’s disease.
Decision Support for Diagnosing Thyroid Diseases Using Machine Learning
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Jayamini, W.K.D.; Weerasinghe, H.D.
Diagnosis of thyroid disorders using two machine learning techniques was studied in this research. Multilayer Perceptron Neural Network with Back-propagation algorithm and Random Forest algorithm were the two algorithms used to build the models for classifying the thyroid diagnosis classes; Hyperthyroidism, Hypothyroidism, Normal. Models were developed with different structures by changing the relevant parameters and the outcomes of the developed models were compared with each other. For developing different neural networks, parameters such as the number of hidden layers, number of neurons in hidden layers and learning rates were changed. For developing different random forest models, parameters such as the number of features per tree and the number of trees in forest were changed. Those models were trained and tested using two different datasets of thyroid diagnosis (Dataset 1 and Dataset 2) which have different attributes that are related to diagnosing thyroid diseases. The models were tested using 10-fold cross-validation while the models were compared and evaluated using the measures Accuracy (%), Mean Absolute Error, Root Mean Squared Error, TP rate, FP rate, Precision and Recall. In diagnosing thyroid disease, both the algorithms performed well. Multilayer Perceptron Neural Network with Backpropagation algorithm performed well for Dataset 1 with an accuracy of 96.7442% and Random Forest algorithm performed well for the Dataset 2 with a mean accuracy level of 98.4915%.
Deep Learning Based Student Attention Monitoring and Alerting System During a Lecture
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Vettivel, N.; Ravindran, V.; Jeyaratnam, N.; Sumathipala, S.
Mindfulness is the ability to fully aware and focuses on the present moment. For students, it is essential to pay full concentration during the lectures. Staying focused while studying is vital for the better performance of any student. In this study, focuses on developing a deep learning-based attention monitoring and alerting system. The proposed system monitors attention of students during a lecture and gives an alert when attention is diverted. The study used mainly three aspects namely Heart Rate Variability, Brain Waves and Facial Expressions to capture the attention level of students while attending a lecture. By using three different aspects, it is expected to overcome the limitations of each aspect. Each aspect is further divided into several parameters, and most significant parameters that respond to the loose of students’ concentration was chosen using principal component analysis to train the deep neural network to measure the students’ concentration level. As the parameters cannot be able to label accurately with concentration, study used an unsupervised learning methodology and it considers the concentration drifting moment as an anomaly and detect it by deducing the pattern of the parameters. When the concentration drops below the threshold system will alert the user. The preliminary experiments reveal how the Facial Expressions, Heart Rate Variability and Brain Waves change with students’ concentration.
Development of Image Processing Algorithm for Vein Detection System
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Wanniarachchi, T.G.; Tharushika, R.T.P.; Panthaka, W.S.P.
The process of obtaining intravenous access, vein puncture is an everyday invasive procedure in medical settings. A major problem faced by the nurses today is difficulty in accessing veins for intravenous drug delivery and other medical situations. Hence a vein detection device which can clearly show veins is a useful biomedical engineering application. The accessibility to existing devices are limited due to their high costs. When considering patients admitted into hospital wards, the nurses have to struggle with majority of them to access a peripheral venous line. The probability of it is as high as 80% depending on the condition of the patient and the location of the hospital. Although a peripheral vein can be accessed in a single attempt, in a substantial number of patients the attending nurse needs multiple attempts to insert the needle successfully. Excessive vein puncture are both time and resource consuming events, which cause anxiety, pain, and distress in patients, or can, lead to severe harmful injuries. Therefore it is a significant problem in emergency rooms and during a hospital stay. This research deals with the design and development of lowcost non-invasive subcutaneous vein detection system based on near infrared imaging. In here our priority is focused for development of image processing algorithm to extract vein pattern from a acquired near infrared image. Vein detection system uses an infrared light source (740 nm) to illuminate veins in hand. A snapshot of the region is taken by the modified visible light camera to IR region and it is subjected to existing image processing techniques and author’s validity function. Finally the extracted vein pattern is used to project back to the skin of the patient.
EduMiner- An Automated Data Mining Tool for Intelligent Mining of Educational Data
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Kasthuriarachchi, K.T.S.; Liyanage, S.R.
Data mining is a computer based information system that is devoted to scan huge data repositories, generate information and discover knowledge. Data mining pursues to find out patterns in data, organize information of hidden relationships, structure association rules and many more operations which cannot be performed using classic computer based information systems. Therefore, data mining outcomes represent a valuable support for decisions making in various industries. Data mining in education is not a novel area but, lives in its summer season. Educational data mining emerges as a paradigm oriented to design models, tasks, methods, and algorithms for exploring data from educational settings. It finds the patterns and make predictions that characterize learners’ behaviors and achievements, domain knowledge content, assessments, educational functionalities, and applications. Educators and non-data mining experts are using different data mining tools to perform mining tasks on learners’ data. There are a few tools available to carry out educational data mining tasks. However, they have several limitations. Their main issue is difficulty to use by non- data mining experts/ educators. Therefore, an automated tool is required that satisfies the data mining needs of different users. The “EduMiner” is introduced to make important predictions about students in the education domain using data mining techniques. R studio, R Shiny, data mining algorithms and several key functionalities of Knowledge Discovery in Databases have been used in the development of “EduMiner”. The functionalities of the tool are very user-friendly and simple for novice users. The user has to configure the tool and provide the appropriate inputs for parameters such as the data set, the algorithms used for mining in advance to obtain the results of the analysis. The pre-processing will be done to clean the data prior to starting the analysis. The tool is capable of performing several analytical tasks. They are; student dropout prediction, student module performance prediction, module grade prediction, recommendations for students/ teachers, student enrollment criteria predictor and student grouping according to different characteristics. Apart from these features, the tool will consist of an intelligent execution of data analysis tasks with real time data as a background service. Finally, the results of the analysis are evaluated and visualized in order to easily understand by the user. Users of education industry can achieve a valuable gain by this tool since, it would be very user friendly to handle and easy to understand the mining results.
Evaluation of Trustworthiness for Online Social Networks Using Advanced Machine Learning
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Mayadunna, H.; Rupasinghe, L.
The trustworthiness of online users has become a current issue in the field of social computing with the rapid popularity of online social networks. The evaluation of trust in social networks has been widely used in situations such as friend – recommendation, e- commerce and access control systems. For sharing and exchanging of information between the trusted users only trustworthiness of the user needs to be determined. One of the key requirements in trust applications is recognizing the trustworthy actors in the network. In the proposed research, a general trust framework will be introduced to calculate the node trust values for social network users by applying machine learning methods. Some selected features of social network are used as the training feature and the measurement whether there is edge between nodes used as label information. Secondly, a training model will be used to calculate the node trust value. Then a recommendation algorithm will be used to calculate node trust score. Finally, the simulation is used to verify the performance of suggested method. For the simulation of experimentation, data from an adaptive social network will be used. The emergence of online social networking (OSN), like Facebook, Twitter, Instagram are allowed people to build and maintain social relationship over the internet. Currently, a large number of users around the globe are connected to the online social networks for sharing and exchanging information. Online social networking is a common platform for communication and sharing different type of information. The popularity has increased of such social networks that have millions of connected users. In online social network, it is important to determine which user gets access to the information related to the user. Information related to trustworthiness of other users can help a user to take decisions about information exchange, sorting and filtering of information. The method will help in building more confidence about using social network among users. Protection of information from untrusted user is crucial aspect in social network. The method enables maintenance of the user privacy and confidentiality by finding trustworthiness of user.
Feature Extraction from Old Tamil Newspapers Using Histogram Minima
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Kasthuri, S.; Darsha, M.; Ranathunga, L.
Archaeological records which provide information about the history of human cultures and past events. Newspapers can be considered as one of the main sources of gathering archaeological data. It can be seen that there exist only a few numbers of systems for the processing of old Tamil newspaper articles. An automated image processing system proposed as a suitable solution to the way of efficient and flexible searching approach, which can be used for old Tamil newspapers. In this paper is presented image processing technique to extract the features such as headlines and sub-headlines from old Tamil newspaper scanned images. Historical newspapers become damaged over time. The images of these newspapers become difficult to read the contents. The quality of the image improved by preprocessing techniques such as grayscale dilation, median filtering, and adaptive binarization. It helps to easily extract needed information on the image. Segment the article and identify the heading of the article will help to improve data manipulation. Feature extraction from old Tamil newspaper images followed these step processes; Horizontal smoothing is necessary to distinguish the paragraphs and empty space between each column; Vertical smoothing is implemented to distinguish between each paragraph and headlines; Logical AND operation combines the outcome of horizontal smoothing and vertical smoothing using AND operation; Height measurement of each block is followed by horizontal projection, that involves scanning of pixels through horizontal arrays to measure the black pixel density against index of each row by using horizontal histogram minima. This step identified horizontals breaking points of individual regions within an article. The four major horizontal regions are headlines, sub-headline, text, and graphics. The irregular block may contain images within texts. Vertical projection can be carried out to distinguish the images among text. In the evaluation process used fifty articles which have different format of paragraph arrangements and also include images. First, identified and got the count of regions manually. After that compared the result from identified regions and got the measurements. The region was identified with articles in the efficiency of 80.09%, headline extraction accuracy was 81.616%.
Finite Element Method based Triangular Mesh Generation for Aircraft-Lightning Interaction Simulation
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Vinotha, K.; Thirukumaran, S.
Lightning is a natural electrical discharge process. Most common lightning strike is Cloud-to-Ground. It occurs when the negative charges accumulated at the bottom of the thundercloud traverse towards the ground to neutralize its charges with the positive earth charges induced due to the thundercloud and electrons travels along the lightning channel. The statistics shows that the commercial aircrafts directly struck by lightning strikes that are under the thundercloud once a year on average. The study of electromagnetic threat due to lightning strikes is important for flight safety and restructuring the aircraft design to mitigate direct lightning effects on the physical material of the aircraft causing damages and indirect effects on the navigation systems in it.The prime objective of this paper is to find the electric field distribution around the aircraft conductor in free space conditions under lightning scenario. For the simulation, the flash of the cloud-to-ground lightning is represented as a wave equation. Finite element method is applied to solve the wave equation for identifying potential distribution and exclusively to electric field calculations. Each of the triangular finite elements are considered and the potential at any nodes within a typical element are obtained. The equation 𝐸 = −𝛻𝑉 represents the relationship between electric potential and electric field which is used to determine the electric field distribution around the aircraft surface by a numerical derivative evaluation technique from the electric potential distribution already obtained. This paper presents an aircraft-lightning interaction simulation under the thundercloud and above the ground by generating two dimensional triangular mesh using finite element method. Significant electric field distribution is observed at the sharp end points of the aircraft. Due to higher radiated electric field, the aircraft-lightning interaction may result in an adverse impact on the aircraft navigation systems and cause damage to its structures. The simulation results would be very useful for studying lightning impact on the aerial vehicles struck by the cloud-to-ground lightning. During the simulation, it was assumed that an aircraft surface is a good conductor and the effects of material properties are left for future studies.
Finite Element Method based Triangular Mesh Generation for Aircraft-Lightning Interaction Simulation
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Vinotha, K.; Thirukumaran, S.
Lightning is a natural electrical discharge process. Most common lightning strike is Cloud-to-Ground. It occurs when the negative charges accumulated at the bottom of the thundercloud traverse towards the ground to neutralize its charges with the positive earth charges induced due to the thundercloud and electrons travels along the lightning channel. The statistics shows that the commercial aircrafts directly struck by lightning strikes that are under the thundercloud once a year on average. The study of electromagnetic threat due to lightning strikes is important for flight safety and restructuring the aircraft design to mitigate direct lightning effects on the physical material of the aircraft causing damages and indirect effects on the navigation systems in it.The prime objective of this paper is to find the electric field distribution around the aircraft conductor in free space conditions under lightning scenario. For the simulation, the flash of the cloud-to-ground lightning is represented as a wave equation. Finite element method is applied to solve the wave equation for identifying potential distribution and exclusively to electric field calculations. Each of the triangular finite elements are considered and the potential at any nodes within a typical element are obtained. The equation represents the relationship between electric potential and electric field which is used to determine the electric field distribution around the aircraft surface by a numerical derivative evaluation technique from the electric potential distribution already obtained. This paper presents an aircraft-lightning interaction simulation under the thundercloud and above the ground by generating two dimensional triangular mesh using finite element method. Significant electric field distribution is observed at the sharp end points of the aircraft. Due to higher radiated electric field, the aircraft-lightning interaction may result in an adverse impact on the aircraft navigation systems and cause damage to its structures. The simulation results would be very useful for studying lightning impact on the aerial vehicles struck by the cloud-to-ground lightning. During the simulation, it was assumed that an aircraft surface is a good conductor and the effects of material properties are left for future studies.
Forecasting Monthly Ad Revenue from Blogs using Machine Learning
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Dias, D.S.; Dias, N.G.J.
Blogs emerged in the late 1990s as a technology that allows Internet users to share information. Since then, blogging has evolved to become a source of living to some and a hobby to others. A blog with rich content and regular traffic could easily be monetized through a number of methods. Affiliate marketing, Google AdSense, offering courses or services, selling eBooks and paid banner advertisements are some of the methods in which a blog could be monetized. There exists, a direct relationship on the revenue that can be generated through any of the above methods and the traffic that the blog gets. Google AdSense is the leader in providing ads from publishers to website owners. All bloggers or blogging website owners who have monetized their blogs, attempt to maximize their revenue by publishing articles in hope that it will generate the targeted revenue. On the other hand, bloggers or blogging website owners that hope to monetize their blog will be greatly benefitted if there was a way to forecast the monthly ad revenue that could be generated through the blog. But there exists no tool in the market that can help the bloggers forecast their ad revenue from the blog. In this research, we are looking at the possibility of finding an appropriate machine learning technique by comparing a linear regression, neural network regression and decision forest regression approaches in order to forecast the monthly ad revenue that a blog can generate to a greater accuracy, using statistics from Google Analytics and Google AdSense. As conclusion, the Decision Forest Regression model came out as the best fit with an accuracy of over 70%
Hybrid Gene Selection with Information Gain and Multi-Objective Evolutionary Algorithm for Leukemia Classification
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Fajila, M.N.F.
Leukemia is a bone marrow cancer with various subtypes such as Acute Myeloid Leukemia and Acute Lymphoblastic Leukemia which require expertise to be identified. Morphological and histological appearances can be used to identify diseases. Yet, precise identification of subtypes is a difficult task. Therefore, subtype detection is a crucial part in prognosis. In this study, a hybrid gene selection approach Information Gain-Multi-Objective Evolutionary Algorithm (IG-MOEA) is proposed to identify Leukemia subtypes. Microarray data consists of thousands of genes where all are not corresponding to disease. Irrelevant and redundant genes have high impact on worst classification performance. Hence, IG is initially applied to preprocess the original datasets to remove irrelevant and redundant genes. Then, further MOEA is used to select a smaller subset of genes for perfect classification of new instances. Gene subset selection highly influences the classification. Further, the subsets selected intern is influenced by the algorithm used for gene selection. Moreover, informative subset of genes can be used efficiently for perfect prediction. Thus, selecting the appropriate algorithm for subset selection is important. Hence, MOEA is used in the proposed study for subset selection. The performance of proposed IG-MOEA is compared against the Information Gain-Genetic Algorithm (IGGA) and Information Gain-Evolutionary Algorithm (IG-EA). Three Leukemia microarray datasets were used to evaluate the performance of the denoted approach. Remarkably, 100% classification was achieved for all the three datasets only with few informative genes using the proposed approach.
The Impact of Soft Productivity Factors on Employee Turnover in IT Industry; A Case Study in Sri Lanka
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Perera, B.A.A.; Rathnayake, R.M.K.T.
Employee turnover has become one of major issues in IT industry in Sri Lanka. Employee turnover is deeply related with the productivity of an organization. Therefore the major purpose of this study is to examine how some selected soft productivity factors affect employee turnover in IT sector and introduce recommendations to reduce turnover rate. Eleven soft productivity factors were considered in the study and they were grouped as workplace environment and employees' capabilities and experience. The major data collection method used in the study was questionnaire survey and descriptive statistics were employed as well. Pilot testing was conducted before data collection and reliability of questionnaire was tested in order to filter the most valid questionnaire. The target population was IT employees who were employed in IT companies at the time. According to the results most employees were totally satisfied with the working place (73.3%) when there were good collaboration with team members (70.4%), no heavy workloads (75.3%), training and development programs (54.2%), when employees were provided with appropriate tools and development resources (70%) and encouraged to rest and refresh (60.9%). Good blend of different characteristic employees had been a motivation to retain in the organization (72.3%). It is concluded that good relationships and collaboration among employees and good blend of different characteristic team members are critical factors which support employee retention. Physical separation of team members has no significant effect on employee turnover if there are sufficient and efficient telecommunication facilities. Comfortable working conditions make employees less stressed and workload does not matter in such environments. When employees’ experience level, skills, and capabilities are higher, they tend to leave organizations for better opportunities.
An Initial Study on Understanding the Effect of Questions Structure on Students' Exam Performance
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Wijesinghe, S.; Irosha, K.P.C.; Rupasinghe, T.
The main challenge in evaluating students’ performance is creating effective assessments which appraises students’ learning rather than their memory power and the practice. According to education theories, creative and carefully designed assessments can clearly evaluate the degree of learning in students. “Scaffolding” which refers to the degree to which a question guides the student through the problem-solving process is a widely used method in aiding students’ learning and conceptual understanding and assessing students’ performance in Science and Technology education. The objective of the current study was to understand the impact of exam question structure on the performance of first year undergraduates specifically focusing on understanding the effect of scaffolded questions. In the current Sri Lankan science education context, there is only a limited number of research studies that are available which provides an insight into the relationship between students’ performance and question features. Current study which was designed to address this issue was conducted as a part of the Chemistry for Technology course at the Faculty of Computing and Technology, University of Kelaniya, Sri Lanka. In this study, two different structures of the same questions were given to students as a part of an in class quiz. First one was a direct question and the second version (scaffolded question) included the same question in a step by step manner and in the latter version, students had to answer several steps to solve the problem. Marks obtained for the two versions were averaged and compared to investigate whether there is any significance of the structure of the questions towards the performance of students. Average mark for the scaffolded question was 82(±20) and the direct question was 71(±35). According to the results, it was clear that the students meet a considerable difficulty in the understanding the direct questions and the scaffolding of questions results in an increase of the performance of students. According to preliminary data, it can be concluded that scaffolding of questions preferentially assist students performance at examinations and surface features such as the structure of the question can play a key role in students’ performance at the examinations. Further studies are currently being conducted to understand whether there is any specific correlation between the improvement in performance as a consequence of scaffolding with the gender, school district and students’ English literacy.
Introducing Novel Classification Methodology to Detect Kidney Disease Patterns in Sri Lanka
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Chathurangi, K.A.A.; Kapila, R.M.; Rathnayaka, T.
The healthcare sector has vast amount of medical data which are not properly analyzed and mined to discover useful information and interesting patterns. Applying data mining techniques on such domain can help medical practitioners to predict even the crucial diseases with ease. This study introduced a novel kidney disease classification methodology in Sri Lankan domain using data mining techniques. Basically there are two types of kidney diseases that can be found in Sri Lanka namely Chronic Kidney Disease (CKD) and Acute Kidney Disease (AKD). The aim of this work is building a model to predict whether a person has a risk on having a kidney disease or not and a model for CKD prediction. The data collected from 108 patients are used to train and test the models. Random Forest algorithm and a multilayered feed forward neural network were used to build the models. Result of this study is a modified Artificial Neural Network with 2 hidden layers to detect kidney disease which gives 0.80952 accuracy and a model with the combination of Random Forest algorithm and Artificial Neural Network with 3 hidden layers for CKD prediction which gives 0.81395 accuracy for testing data. The constructed models give high accuracy and minimum error rate when comparing with the other data mining algorithms.
Investigation of the Degradation Processes Effect on the Properties of the Industrial Cutting Tool used in Packaging Process
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Voloshyna, A.; Romaniuk, S.
The paper investigates the metal structure carbide inhomogeneity of the cutting tool made from high-alloy steel, used in industrial process of packaging candies, after its service life period. The chemical composition of steel is current: C - 2.2%, Si - 0.4%, Mn - 0.35%, Cr - 12.12%, Fe - 84.8%, Mo - 0.12%. The increased content of carbon and chromium leads to the formation of an amount of special doped carbides in the composition structure. Accordingly to the X-ray diffraction analysis, it was detected that the type of carbides conforms to Cr7C3. The amount of carbides and their size were determined with the computer program ThixometPro. As indicated by the metallographic analysis of the separate zones of the tool, the size and the number of special doped carbides differ in the images of the metal structure. Therefore, the structure of the middle part and at the edge of the operating surface were comparatively analyzed. The total amount of carbides in the middle part of the tool structure equals 14.4% of the metal matrix and reaches 8.15% at the edge of the operating surface. The structural inhomogeneity and the presence of large doped carbides were detected in the middle zone, wherein, the share of small carbides is 20.8% of the total volume of the carbide phase. There is a lack of large special carbides and the area of 69.2% carbides is not exceeding 4.75 µm at the distance up to 100 µm from the edge of the working surface. Moving further from the edge, the area and volume of carbides increase. The carbide inhomogeneity along the cross-section occurs as a result of doped carbides crushing under the stresses action during the service life. From the working surface edge to the depth dispersed carbides are lining up at an angle of 45º, forming centers of crack initiation. In virtue of the analysis, it is recommended to apply an additional hardening by the PVD method to stabilize the operating surface layer under the deformation.
Investigation of the Impact of Clay as a Bulking Agent for Food Waste Composting at a Controlled Raised-up Temperature
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Jayawardana, M.D.S.B.; Milani, Y.; Silva, C.D.; Wijesinghe, S.
In agriculture, the nourishing and substantial quality of soil can be upgraded through transfiguring organic matter in food waste into humus like substance, which is called food waste composting. This is very important as food waste leads to cause odor and pollute the environment. The moisture content (MC), nitrogen content, C/N ratio and aeration in the compost material can be altered through various bulking agents used during the process. The usage of these bulking agents enhances the biodegradation of food waste and transformation of it into effective compost. Therefore, the entire composting process relies on the indispensable role of the bulking agents. Thus, this study was ultimately aimed to evaluate the influence of using clay as the bulking agent for food waste composting at a controlled high temperature (500C). Here a controlled raised-up temperature was used to lead rapid activation of thermophilic microbes. A consecutive five-day study was carried out to analyze the fluctuations of PH, MC and organic matter content (OMC) by preparing composting feedstock using clay as the bulking agent in four different weight percentages (0%, 5%, 10% and 25%). Using a Scanning Electron Microscope (SEM) surface morphology of the samples was analyzed at the initial stage and after five days composting. The analysis of physical parameters was evident that the organic matter was effectively converted to compost at 500C as all the parameters followed the corresponding gradual fluctuations which are presented at the quality compost production. According to the results, no effect was found from clay to control the PH of the composting process of food waste samples. With the increasing of clay percentage there was no significant change of PH was noticed compared to the blank waste sample. With the increment of the clay percentage of the composting feedstock, initial MC was dropped. Furthermore, by the increasing of the clay content of the samples MC was highly reduced. Similarly, OMC was also drastically decreased with the upswing of clay percentage. According to the observations, it can be concluded that clay has been acted as a good bulking agent to food waste composting. At this elevated temperature Food waste composting process had shown a significantly improvement. Presently, further studies are being carried out to further optimize the percentage of clay for food waste composting process at elevated temperature.
Mobile Biometrics: The Next Generation Authentication in CloudBased Databases
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Bhatt, C.; Liyanage, S.R.
In this period of data innovation, cell phones are generally utilized around the world for fundamental correspondences, as well as an apparatus to manage anyplace, whenever data. These situations require a high security level for individual data and protection assurance through individual distinguishing proof against un-approved use if there should be an occurrence of robbery or fake use in an organized society. At present, the most received technique is the check of Personal Identification Number (PIN), which is risky and won't not be anchored enough to meet this prerequisite. As is represented in a review (Clarke and Furnell, 2005), numerous cell phone clients view the PIN as badly arranged as a secret key that is sufficiently confounded and effortlessly overlooked and not very many clients change their PIN frequently for higher security. Subsequently, it is liked to apply biometrics for the security of cell phones and enhance dependability of remote administrations. As biometrics intends to perceive a man utilizing special highlights of human physiological or conduct attributes, for example, fingerprints, voice, confront, iris, stride and mark, this verification technique normally gives an abnormal state of security. Expectedly, biometrics works with particular gadgets, for instance, infrared camera for securing of iris pictures, increasing speed sensors for step obtaining and depends on expansive scale PC servers to perform ID calculations, which experiences a few issues including massive size, operational many-sided quality and greatly surprising expense. Adding a wireless dimension to biometric identification provides a more efficient and reliable method of identity management across criminal justice and civil markets. Yet deploying cost-effective portable devices with the ability to capture biometric identifiers – such as fingerprints and facial images – is only part of the solution. An end-to-end, standards-based approach is required to deliver operational efficiencies, optimize resources and impact the bottom line. While the use of mobile biometric solutions has evolved in step with the larger biometrics market for some time, the growing ubiquity of smartphones and the rapid and dramatic improvements in their features and performance are accelerating the trend. This is the right time to take a closer look at mobile biometrics and investigate in greater depth how they can be used to their potential. Consolidated with cutting edge detecting stages can identify physiological signals and create different signs, numerous biometric strategies could be executed on phones. This offers an extensive variety of conceivable applications. For example, individual protection assurance, versatile bank exchange benefit security, and telemedicine observation. The utilization of sensor information gathered by cell phones for biometric ID and verification is a rising boondock that must be progressively investigated. We review the state-of-the-art technologies for mobile biometrics in this research.
Mobile Telecommunication Customers Churn Prediction Model
(3rd International Conference on Advances in Computing and Technology (ICACT ‒ 2018), Faculty of Computing and Technology, University of Kelaniya, Sri Lanka., 2018) Chathuranga, L. L. G.; Rathnayaka, R.M.T.B.; Arumawadu, H.I.
The present Sri Lankan mobile industry is extremely dynamic, with new services, technologies, and carriers constantly altering the landscape. Then customers have more choices. So, Predict customer churn is one of the most challengeable target in the telecommunication industry today. The major aim of the study is develop a customer churn prediction model by considering some soft factors like monthly bill, billing complaints, promotions, hotline call time, arcade visit time, negative ratings sent, positive ratings sent, complaint resolve duration, total complaints, and coverage related complaints. This study introduces a Mobile Telecommunication customer churn prediction model using data mining techniques. In this study, three machine learning algorithms namely logistic regression, naive bayes and decision tree are used. Indeed, twenty attributes are mainly carried out to train these three algorithms. Furthermore, the back propagation neural network was trained to predict customer churn. Data set used in this study contains 3,334 subscribers, including 1,289 churners and 2,045 non-churners. According to the results, the trained neural network has two hidden layers with 25 total neurons. The proposed Artificial Neural Network result gives 96% accuracy for mobile telecommunication customer churn prediction. The estimated results suggested that the proposed algorithm gives high performances than traditional machine learning algorithm.