Statistics & Computer Science
Permanent URI for this collectionhttp://repository.kln.ac.lk/handle/123456789/3751
Browse
26 results
Search Results
Item HOURLY SOLAR RADIATION FORECASTING USING ARTIFICIAL NEURAL NETWORK MODEL FOR COLOMBO, SRI LANKA(Advances and Applications in Statistics, 2019) Saumyamala, M.G.A.; Chandrasekara, N.V.Sri Lanka is a tropical country located close to the equator with abundant sunlight throughout the year. For efficient utilization of this solar resource for power generation in photovoltaic (PV) systems and agricultural modelling, prior knowledge of global solar radiation (GSR) in the future is important. Limited availability of onsite GSR data and the high cost are the main barriers in forecasting GSR for Sri Lanka. As a solution this study suggests an artificial neural network (ANN) model to forecast hourly solar radiation using weather data and solar angles to forecast GSR in Colombo, specifically using feedforward neural network (FFNN) trained with Levenberg- Marquardt (LM) back propagation algorithm. Hourly weather data for 6 weather variables and two solar angles from 1st of March 2017 to 14th of February 2018 were used for training, validation and testing the network. Input parameters and training parameters were adjusted to identify the most accurate network configuration and the performance of the network was measured using normalized mean squared error (NMSE). Coefficient of determination (R2) measured to identify the appropriateness of using weather variables and solar angles to forecast solar radiation. The final hourly FFNN model consists of 2 hidden layers and there are 5 neurons and 3 neurons in each layer respectively. This model was able to forecast hourly solar radiation with 0.0961 NMSE and the R2 was 90.39%. This implies the capability of this model for prediction of global solar radiation when unseen weather data input supply to the model and ensure the accuracy of the result.Item Factors associated with induction of labour and pregnancy outcomes in 14 healthcare facilities in Sri Lanka(Journal of Science 2019, Department of Statistics & Computer Science, Faculty of Science, University of Kelaniya, Sri Lanka, 2019) Gunawardana, J. R. N. A.; Perera, S. S. N.Induction of Labour (IOL) is an important practice that is carried out commonly in modern day obstetrics. In medium to large healthcare facilities in Sri Lanka, it is estimated that approximately 35.5% of all deliveries involve IOL. This research attempts to identify the factors that affect IOL and to assess the association between IOL and the pregnancy outcome. In this study, we considered 18309 women who were admitted to 14 healthcare facilities for delivery in 3 randomly selected provinces in Sri Lanka (Western, Southern and Eastern provinces), during July to October 2011. Multinomial Logistic Regression model (MLR) and Fuzzy Expert System (FES) were used to identify the factors that lead to IOL. The MLR model predicts IOL with a classification rate of 65.5% and the FES predicts IOL with an accuracy of 55.10%. 1Maternal age, number of previous caesarian sections,number of previous births, estimated gestational age, Pre-Eclampsia, number of previous pregnancies, Placenta Preavia, Abruption Placenta, total number of neonates delivered, birth weight and Maternal Severity Index (MSI) were identified as factors associated with IOL. Neonatal status after seven days of life can also be predicted using the developed FES. FES is predictive of IOL and birth outcome, where if the FES score is between 0.8570 and 0.8854, the patient will belong to the induced group and the baby would be alive after seven days of birth. This study concludes that, MLR and FES models can be used to predict IOL outcomes. These findings can be informative to healthcare providers when counselling women for labour induction and develop evidence-based protocols on IOL.Item An Ensemble Technique For Multi Class Imbalanced Problem Using Probabilistic Neural Networks(Advances and Applications in Statistics, 2018) Chandrasekara, N.V.; Tilakaratne, C.D.; Mammadov, M.A.The class imbalanced problem is one of the major difficulties encountered by many researchers when using classification tools. Multi class problems are especially severe in this regard. The main objective of this study is to propose a suitable technique to handle multi class imbalanced problem. Probabilistic neural network (PNN) is used as the classification tool and the directional prediction of Australian, United States and Srilankan stock market indices is considered as the application. We propose an ensemble technique to handle multi class imbalanced problem that is called multi class undersampling based bagging (MCUB) technique. This is a new initiative that has not been considered in the literature to handle multi class imbalanced problem by employing PNN. The results obtained demonstrate that the proposed MCUB technique is capable of handling multi class imbalanced problem. Therefore, the PNN with the proposed ensemble technique can be used effectively in data classification. As a further study, other classification tools can be used to investigate the performance of the proposed MCUB technique in solving class imbalanced problems.Item Homogenization of Daily Temperature Data(American Meteorological Society, 2017) Hewaarachchi, A.P.; Li, Yingbo; Lund, Robert; Rennie, JaredThis paper develops a method for homogenizing daily temperature series. While daily temperatures are statistically more complex than annual or monthly temperatures, techniques and computational methods have been accumulating that can now model and analyze all salient statistical characteristics of daily temperature series. The goal here is to combine these techniques in an efficient manner for multiple changepoint identification in daily series; computational speed is critical as a century of daily data has over 36 500 data points. The method developed here takes into account 1) metadata, 2) reference series, 3) seasonal cycles, and 4) autocorrelation. Autocorrelation is especially important: ignoring it can degrade changepoint techniques, and sample autocorrelations of day-to-day temperature anomalies are often as large as 0.7. While daily homogenization is not conducted as commonly as monthly or annual homogenization, daily analyses provide greater detection precision as they are roughly 30 times as long as monthly records. For example, it is relatively easy to detect two changepoints less than two years apart with daily data, but virtually impossible to flag these in corresponding annually averaged data. The developed methods are shown to work in simulation studies and applied in the analysis of 46 years of daily temperatures from South Haven, Michigan.Item Determining and Comparing Multivariate Distributions: An Application to AORD and GSPC with their related financial markets(2016) Chandrasekara, N.V.; Mammadov, M.; Tilakaratne, C.D.Many real world applications are associated with more than one variable and hence, identifying multivariate distributions associated with real world problems portrays great importance today. Many studies can be found in the literature in this aspect and most of them are associated with two variables/dimensions and the maximum dimension of multivariate distribution found in the literature is four. Different optimization techniques have been used by researchers to find multivariate distributions in their studies. Numerical methods can be identified as more preferable than analytical methods when the dimension of the problem is high. The main objective of this study is to identify the multivariate distribution associated with the return series of Australian all ordinary index (AORD) and those of the related financial markets and compare it with the multivariate distribution of return series of the US GSPC index and its related financial markets. No research were found in the literature which were aimed at finding aforesaid multivariate distribution and comparisons. Moreover no evidence found for identifying a multivariate distribution with six dimensions. Five financial markets: Amex oil index, Amex gold index, world cocoa index, exchange rate of Australian dollar to United States dollar and US GSPC index were found to be associated with AORD. Hence the attempt was to derive the multivariate distribution of return series of AORD and these five return series and therefore the optimization problem of the study is a six dimension problem which associated with forty three parameters need to be estimated. A local optimization technique and a global optimization technique were used to estimate the parameters of the multivariate distribution. Results exhibit that the parameter estimates obtained from the global optimization technique are better than the parameter estimates obtained from the local optimization technique. The multivariate distribution of return series of AORD and related financial markets is central, less peaked and have fat tails. A comparison was done with another multivariate distribution of a return series of a leading stock market index: GSPC and return series of its associated financial markets and found that both distributions are alike in shape. Two periods were identified in the AORD series and found that the shape of the multivariate distribution of one period is similar to the shape of the multivariate distribution of full data set while the shape of the multivariate distribution of the other period is dissimilar to that of full data set.Item An Application of 5-fold Cross Validation on a Binary Logistic Regression Model(2016) Attanayake, A.M.C.H; Jayasundara, D.D.M.; Peiris, T.S.G.Abstract Internal validation techniques can be used to check the predictive ability of the developed models. The most common internal validation techniques are split sample methods, cross validation methods and bootstrapping methods. The split sample methods are inefficient with the small size of data sets. The bootstrapping methods are efficient with the knowledge of computer programming languages. The cross validation methods are not very popular in practice. Therefore, in this study 5-fold cross validation method of cross validation techniques is applied to validate the predictive ability of a binary logistic regression model. The binary logistic regression model was fitted on a data set of UCI machine learning repository. Results of the cross validation reveal that low value of optimism and high value of c-statistic in the fitted regression model indicate an acceptable discrimination power of the developed model.Item An android based live tour guide for Sri Lanka(Sri Lanka Association for the Advancement of Science, 2016) de Silva, A.D.; Liyanage, S.R.Item Estimating parameters of multivariate scaled t distribution of GSPC and its associated financial indices(2015) Chandrasekara, N.V.; Mammadov, M.A.; Thilakaratne, C.D.Item Identifying distributions of selected stock returns(2015) Chandrasekara, N.V.; Thilakaratne, C.D.; Mammadov, M.A.Item The potential of brain computer interfacing for sustainable development(National Science foundation, 2016) Liyanage, S.R.
- «
- 1 (current)
- 2
- 3
- »