http://ir.lib.seu.ac.lk/handle/123456789/5450 2026-04-06T08:19:58Z http://ir.lib.seu.ac.lk/handle/123456789/5460 Title: Common practices and abundance of insects in rice mills in Polonnaruwa District of Sri Lanka Authors: Dissanayaka, D. M. S. K.; Wijayaratne, L. K. W. Abstract: The facilities and conditions available in rice mills have a direct impact on the quantitative and qualitative output of rice production. Despite the great contribution towards the national rice production from the Polonnaruwa district, no recent updates on the existing conditions of rice mills in this area are available. Insect infestation in feed mills is reported from other countries but such information in Sri Lanka is minimum. The current study survey investigated the existing condition of the rice milling industry and insect infestation in mills in Polonnaruwa district. The mills were randomly selected and information on the milling process is collected. The samples of milling output were collected from different locations of the mills and analyzed for the presence of insects and their diversity. Large, medium, and small-scale rice mills were found. The milling and storage capacity and presence and diversity of insects varied with the type of mill. The current study provides insights on the existing status of the rice milling industry and the abundance of insects in such facilities in the Polonnaruwa district. This information would be important for decision making to augment the efficiency in rice milling and to design appropriate protective methods safe guard the milling output from insect infestations. 2020-08-01T00:00:00Z http://ir.lib.seu.ac.lk/handle/123456789/5459 Title: Decision tree based automated prediction of infant low birth weight Authors: Akmal Jahan, M. A. C.; Razmy, A. M. Abstract: Data engineering with decision trees plays a vital part in the field of health and medical diagnosis. Low birth weight (LBW) is the single most important factor determining the survival chances of an infant and predicting this LBW based on the maternal factors through data engineering can be another contribution to the medical diagnosis. In this research, decision tree classifiers are used to predict the incidence of low birth weight of newborns based on maternal factors. A set of decision tree classifiers such as C4.5, Random Tree, Random Forest, Decision Stump, Logistic Model Tree, REP Tree (reduced error pruning tree) and BF (best-first) trees are assessed for this classification purpose. These classifiers are evaluated for their accuracy and time complexity in classifying the childbirth weight. A set of data collected from pregnant mothers in the hospitals located in the Eastern part of Sri Lankan have been used for this study. From the experimental analysis, Random Forest produces the highest accuracy as 79.71 % while BF Tree and C4.5 show 79.23%. However, for the time complexity, Random Tree, REP Tree and C4.5 perform less than 1 second whereas the Random Forest utilizes around 5 seconds. On the other hand, BF Tree and C4.5 show precision and recall as 0.792 and 1.0 respectively. Overall, C4.5 could be the best classifier to construct a decision tree model for the prediction of LBW with the acquired data set. 2020-12-01T00:00:00Z http://ir.lib.seu.ac.lk/handle/123456789/5458 Title: Influence of grain fractions on the progeny adult emergence of the lesser grain borer, rhyzopertha dominica (F.) (coleoptera: bostrychidae) in stored rice varieties and pulses Authors: Hettiarachchi, J. M.; Dissanayaka, D. M. S. K.; Wijayaratne, L. K. W. Abstract: Infestation of stored grains by the lesser grain borer Rhyzopertha Dominica causes serious quantitative and qualitative losses. Grains get fractioned during milling and the availability of brokers with the whole grains may alter the rate of insect infestation and progeny adult emergence. Traditional rice varieties are being reconsidered in addition to the new improved varieties in the diet of Sri Lankans due to the accompanying health benefits. In this study, the progeny development of R. Dominica in locally available traditional and new improved rice varieties and pulses containing different percentages of grain fractions was determined at 2-week intervals following initial infestation. More progeny adults emerged in rice than on pulses. In rice, Red Kekulu and Kuruluthuda varieties had higher adults emerged than in other varieties. Blackgram and green gram had the lowest adult emergence. Increased percentage of grain fragments in the sample decreased the progeny adult emergence. More R. Dominica adults emerged later than early stages following initial infestation. Minimization of initial infestation reduces damage caused by R. Dominica. The future studies need to test how the infestation by other insect species is affected by grain type and availability of broken grain in the medium to ensure better protection of stored grains from insect damage. 2020-12-01T00:00:00Z