000			04706ntm a2200373 i 4500
999			_c99338 _d99344
003			MY-KuUP
005			20251125110726.0
006			t||||fr|||| 000 0
007			ta
008			230410t20222022my a|||fr|||| 000 0 eng d
020			_aTHE0009584 (Local) _qHardback
040			_aUMP _beng _cUMP _erda
090			_aFTKPM .S89 2022 r Thesis
100	1		_aSuziyanti Binti Zaib, _eauthor.
245	1	0	_aThe study of raw water based on quality parameter using smell-print sensing device / _cSuziyanti Binti Zaib
264		1	_aKuantan, Pahang : _bUMP, _c2022
264		4	_c©2022
300			_axiv, 122 pages : _billustrations (some color) ; _c30 cm. + _e1 CD-ROM
336			_2rdacontent _atext
336			_2rdacontent _atext
337			_2rdamedia _aunmediated
337			_2rdamedia _acomputer
338			_2rdacarrier _avolume
338			_2rdacarrier _acomputer disc
347			_2rda _atext file _bPDF
500			_aFaculty of Manufacturing and Mechatronic Engineering Technology
502			_aThesis (Master of Science) -- Universiti Malaysia Pahang – 2022
504			_aIncludes bibliographical references
520	3		_aWater is a renewable natural resource, and its quality plays an essential role in assessment, monitoring, and management. The two primary water sources in Malaysia are groundwater and surface water, used for daily activities, drinking, domestic and industrial purposes. Cleanliness and purity of drinking water are the essential requirements for human health worldwide. Thus, it is important to know the water body source content and water quality to minimize risks to health during consumption. Water quality degrades due to numerous factors such as inorganic contaminants, heavy metals, microbial contaminants, and contaminants at the water source. Hence, humans need to be concerned about this issue and monitor the quality so that the water is safe to drink, not harm the human's body. All drinking water sources must meet the required quality standard. Nowadays, researchers pay great attention to water quality due to the high demand for clean water and population growth. One of the available methods to monitor the quality is using a spectrophotometer. This method needs more procedures, complicated, not suitable for onsite, and needs the expertise to handle the device. Therefore, this research aims to establish a case library profile for groundwater samples based on smell-print, classify several cases of water quality using Case-Based Reasoning (CBR) and K-Nearest Neighbor (KNN), and evaluate the performance measure based on the classification model formulated in CBR and KNN. To fulfill the objectives, an instrument that mimics the human nose, E-Anfun, is used because of its in-situ, less complicated and friendly-odor related device to use in this research. The samples were prepared based on three important parameters; Iron-Fe, Fluoride-F, and pH in the water laboratory plant. The preparation of samples is based on a set of guideline standards recommendation by the Ministry of Health. The data collection of water samples is taken five times for each sample with the estimation of time about two minutes per experiment. The software used to store the data is Microsoft Excel and then proceeded with OCTAVE and MATLAB software. The raw data is then restructured to rescale the value to the range of 0-1 by applying the normalization technique. The size of the dataset of the collected samples was minimized by mean calculation in features extraction that was also used as an input assignment for CBR and KNN classification. CBR consists of four cycles: retrieve, reuse, revise, and retain is used in performing the intelligent classification by solving a new problem based on the successful solution of the previous case. In the meantime, K-Nearest Neighbor (KNN) was used to enhance the CBR by classify the data sample based on learning data that is located closest to the object. In other words, CBR and KNN are the methods used in this research to classify water quality The evaluation of the CBR and KNN classification were measured using recognized confusion matrix. The finding results of CBR for ten samples show that the accuracy is 98.533 %, sensitivity is 93.366 % and specificity is 99.171 %. As for KNN with K=3, the performance rate is 98.500 % of accuracy, 92.691 % of sensitivity and 99.171 % of specificity. Both CBR and KNN performance measures' successful achievements indicated that groundwater quality odor profile is classifiable.
610	2	0	_aFaculty of Manufacturing and Mechatronic Engineering Technology _xDissertations
650		0	_aUniversities and colleges _xDissertations
650		0	_aTheses
942			_2lcc _cTHESIS