Browsing by Author "Kumara, B.T.G.S."

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Data Mining Approach for Landslide Prediction Using Support Vector Machine for Rathnapura District, Sri Lanka
    (Uva Wellassa University of Sri Lanka, 2019-02) Madawala, C. N.; Kumara, B.T.G.S.; Indrathilaka, L.
    Haphazard development activities on mountain slopes and inadequate attention to construction aspects have led to the increase of landslide and consequently sustaining damages to lives and infrastructure. Nearly 3275 sq.km of area spread over the Rathnapura District, seems to be highly prone to land sliding and mass wasting of 2178 sq.km. Landslides occurred in many regions of Rathnapura district Eheliyagoda, Ayagama, Kalawana, and Nivithigala DS divisions, and nearly 90 deaths have reported according to National Research Building Organization (NBRO) 2017 records. Most landslides or potential failures could be predicted fairly and accurately if proper investigations were performed in time. The primary objective of this study is landslidehazard mapping and risk evaluation to determine the real extent, timing, and severity of landslide processes in Rathnapura district, where such knowledge will provide the most significant benefit to government officials, consulting engineering firms, and the general public in avoiding the landslide hazard or in mitigating the losses. Data mining approach can be used to develop prediction models using existing data. Support Vector Machine (SVM) was selected for this study to possess a strong capability to predict landslides by causative factors, slope, land use, elevation, geology, Soil Materials and triggering factor; rainfall was extracted and applied to the SVM. This research introduces a methodology to produce a more relevant and accurate prediction of the landslide and identify the relationship between the hydrological characteristics, soil characteristics and the landslide vulnerability within the study area. Moreover, an improvement of the hazard monitoring, accuracy of early warning and disaster mitigation was performed. The SVM procedure was found that all of the factors had relatively positive effects on the landslide. Based on these results indicate that SVMs can be useful and practical for landslide susceptibility analysis.
  • No Thumbnail Available
    Item
    Is Rainfall Prediction Model Tested in One-Time Point Sufficient?
    (Uva Wellassa University of Sri Lanka, 2020) Dananjali, K.T.; Ekanayake, J.B.; Karunaratne, A.S.; Kumara, B.T.G.S.
    Many rainfall predictions models have been proposed. The common methodology followed by those models is that the model is trained using the data before the target and tested the model in one or a few points and claimed that the model is generalized. However, this project shows that the above procedure is not sufficient to generalize a rainfall prediction model as in some target periods the models failed to achieve a decent prediction quality. The models such as Multilayer Perceptron (MLP), M5P, and Linear Regression-were trained from the weather data collected between the years 2002 and 2015 from the station located at Badulla, Sri Lanka. Initially, the target period was set in the last week of the dataset and the training period was one week before the target week. Then, the training period was extended by one week, until the maximum length of the training period reached, keeping the target fixed. Next, the target period was brought back one week and the same procedure was repeated resulting in 695 models. The prediction quality was measured using Mean Absolute Error (MAE) and represented in heat-maps. The heat-maps show that the prediction quality varies over time. The highest accuracy was given by the MLP so that the MAE has fallen between 0 and 10 mm in 61.7% of the total instances. This indicates that testing models in one or a few time points are not sufficient for the generalization. Further, the reasons for such drastic changes in prediction quality will be investigated in our future projects. Keywords: Linear regression, M5P, Prediction, Multilayer perceptron, Rainfall
  • No Thumbnail Available
    Item
    Social Media and Online News Analytics for Identifying Crime Patterns in Crime Prediction
    (Uva Wellassa University of Sri Lanka, 2020) Sandagiri, S.P.C.W.; Kumara, B.T.G.S.; Banujan, K.
    Social media provides opportunities for users to share their thoughts freely. Every year they generate a large volume of data. In the context of social media, they may include hidden details, which may convey significant events. Crime prediction with the help of Social media provides new dimensions in researches. This research aims to collect data from Twitter posts and validate them using online news to avoid false data. First and foremost, we selected the top crimes happening in the world after an extreme literature review. We used Twitter API and News API to fetch data from Twitter and News blogs. We used two filters to collect data. In the first filter, we fetch Twitter posts and News posts for a specific time duration. These data are fetched by using keywords that relate to crime. In the second filter, eliminate noisy Twitter posts from the collected dataset. We have collected many noisy posts in both sources, i.e. Twitter and News. With the help of collected datasets, we will compare each tweet and news datum and give ratings for comparison data. We can build a crime prediction model with integrating data. The result shows that 68% of collected Twitter posts are excluded after using the second filter. Future development can divide into two main parts. To get more accuracy, we can integrate other factors that affect crime prediction such as weather, human behavior analysis data and we can improve the second filter using the SVM algorithm. Secondly, we can integrate other Social media platforms to fetch data. Keywords: Crime prediction, Social media, Twitter, News