http://dlib.iit.ac.lk/xmlui/handle/123456789/662 2026-04-08T18:10:33Z 2026-04-08T18:10:33Z Alahakoonge, Rivini Hansika Pramodi http://dlib.iit.ac.lk/xmlui/handle/123456789/674 2021-08-05T05:03:29Z 2020-01-01T00:00:00Z

Implementing Ontology-Based Chatbot for Enterprises Resource Planning Systems Alahakoonge, Rivini Hansika Pramodi Over the years Enterprises Resource Planning (ERP) systems have changed certain features about their products. They have made it easy to use. But on the other hand, it was not always user friendly due to complex interfaces and business rules. Users must complete a set of training to use the complex interfaces and follow the predefined business rules. Mainly ERP systems work with millions of records daily. Therefore, with these data-intensive modules, it is important to give meaningful insights fast and accurately. By considering data volume, and potential business opportunities, this thesis presents novel solutions to use the predefined well-structured knowledge as an Ontology and communicates with users in natural language using a chatbot. The chatbot is the interface that uses the to communicate with users. In this sense, the proposed solution has come up with Ontology-Based Chatbot for Enterprises Resource Planning (OBC-ERP) framework for the selected company.

2020-01-01T00:00:00Z Seneviratne, Dinindu http://dlib.iit.ac.lk/xmlui/handle/123456789/673 2021-08-05T04:55:56Z 2020-01-01T00:00:00Z

A Valuation Model to Derive Land Prices in Sri Lanka Seneviratne, Dinindu Real Estate is known to be a scarce resource in the modern world. Lands which can be categorized as a subset of real estate refers to the area on earth which people use to build buildings for residential or business use. With the drastically increasing population and infrastructure development, lands and their availability are drastically reducing. The supply of land as a resource is generally below that of the demand for lands. Hence public are ready to pay higher amounts to buy lands. This study focuses on developing a theoretical model for valuation of lands based on each land’s significant factors. Simple Linear Regression was used to understand the relationship between the price of lands and the variability if significant factors. Then Multiple Linear Regression was used to build a linear model to value the lands. Data from an online web portal and a public survey was used to train and build the model. Due to many dimensions being available, a linearity check was conducted to select the most correlated factors out of the group of factors collected. A descriptive study was done in parallel to understand the distribution of data in the variables. The multiple linear regression model was trained using the reduced dimensions and an equation was formed which derives the price of a land when the values for variables were given. The model trained showed a considerable RMSE value

2020-01-01T00:00:00Z Salley, Abdul Qadir Reyyaz http://dlib.iit.ac.lk/xmlui/handle/123456789/672 2021-08-05T04:49:03Z 2020-01-01T00:00:00Z

Raw Material Risk Prediction Tool Salley, Abdul Qadir Reyyaz Apparel Manufacturing is considered to be one of the most important sectors in a country like Sri Lanka especially when the country’s economy depends on exports. With the introduction of Industry 4.0, the Apparel manufacturing sector is moving towards data driven decision making culture in order to improve the quality and cut down costs in order to be competitive. Due to the sheer volume of the daily transactions processed on a daily basis over multiple data platforms, the apparel manufacturing sector data will be important to derive meaningful insights that will enable to make proactive decisions to improve the performance of the sector. Material Risk is a significant aspect in terms of cutting down the lead time and minimizing unnecessary expenditures in the apparel manufacturing industry. The apparel manufactures will have to bear the cost if there are is any reordering materials due to material issues. In the current context there is not much solutions available for the apparel to predict the material risk beforehand. Hence, introducing a data driven material risk prediction solution that could beneficial to the apparel manufacturing industry. When creating a data driven solution, it is mandatory to have an understanding of the Apparel Manufacturing domain. Considering the fact, there is only a limited number of researches that has focused on Apparel Manufacturing there is less concepts published, this study has used a dataset from the ERP of a leading apparel company in Sri Lanka. This was done to get a real-life scenario of an Apparel Manufacturing organization to predict the martial risk based on past experiences of the company. The data was pre-processed using statistical concepts to ensure accuracy and mitigate the biasness of the dataset. Concepts such as logistic regression and decision trees have been used mainly for this study. The first time through percentage was assessed in two instance and was broke down to offer methods for improving first time through percentage activities with the utilization of data. The research determines a prediction of materials to improve supply chain in the apparel manufacturing industry.

2020-01-01T00:00:00Z Senevirathna, Dewagiri Mudiyanselage Gaya Ramya http://dlib.iit.ac.lk/xmlui/handle/123456789/671 2021-08-03T17:55:23Z 2020-01-01T00:00:00Z

Predictive Model for Analyzing Relationship Among Bank Profitability Bank Specific Macroeconomic Determinants and Elements of Expected Rate of return: Evidence from Sri Lanka Banking Sector Senevirathna, Dewagiri Mudiyanselage Gaya Ramya

2020-01-01T00:00:00Z