Comparing euclidean distance and nearest neighbor algorithm in an expert system for diagnosis of diabetes mellitus

2020 ◽  
Vol 30 ◽  
pp. 374-377 ◽  
Author(s):  
Reza Zubaedah ◽  
Fransiskus Xaverius ◽  
Hasanudin Jayawardana ◽  
Serli Hatul Hidayat
2021 ◽  
Vol 54 (6) ◽  
pp. 421-424
Author(s):  
H. Kim ◽  
D. A. Chuvikov ◽  
D. V. Aladin ◽  
O. O. Varlamov ◽  
L. E. Adamova ◽  
...  

2011 ◽  
Vol 63-64 ◽  
pp. 264-267
Author(s):  
Hong Zhi Liu ◽  
Zhen Hua Wang

To improve efficiency of information engineering surveillance, expert system was introduced. First, describes the theory of expert system and the necessity of expert system for information engineering surveillance; Second, designed the system architecture, analysis how to store the cases of the system and describes a case retrieval algorithm, introduce information entropy, compute information gain; Finally, make an better Nearest neighbor algorithm method, the improved algorithm has overcome the traditional K-NN method’s inadequate.


2020 ◽  
Vol 7 (2) ◽  
pp. 417
Author(s):  
Ikhsan Wisnuadji Gamadarenda ◽  
Indra Waspada

<p class="Abstrak">Penyakit ginjal kronis (PGK) merupakan masalah kesehatan publik di seluruh dunia dengan insiden yang terus meningkat. Berdasarkan sumber dari BPJS Kesehatan, perawatan PGK merupakan ranking kedua pembiayaan terbesar setelah penyakit jantung. Pendeteksian PGK juga memerlukan banyak atribut sehingga membutuhkan biaya yang cukup mahal. Oleh sebab itu dibuat sistem dengan tahapan data mining berbasis web yang memudahkan untuk melakukan deteksi PGK, sehingga PGK dapat dicegah, ditanggulangi, dan kemungkinan mendapatkan terapi yang efektif lebih besar jika diketahui lebih awal. Proses penelitian ini menggunakan sebuah rangka kerja<em> data mining</em> <em>Knowledge Data Discover</em>y (KDD). Dalam skenario rangka kerja yang digunakan, sistem ini menggunakan Algoritme <em>Backward Elimination</em> untuk mengurangi jumlah atribut yang dipakai dengan tujuan untuk mengurangi jenis pemeriksaan yang dilakukan, dan Algoritme k-<em>Nearest Neighbor</em> sebagai algoritme klasifikasi untuk mendeteksi penyakit. Hasil pemodelan terbaik <em>data mining</em> dari sistem yang dibuat menggunakan <em>Backward Elimination</em> (α = 0,05) dan kNN (<em>k = </em>3) dengan pertimbangan penurunan biaya pemeriksaan dan sensitivity tertinggi. Rekomendasi sistem menghasilkan 10 atribut yang terpilih dari 24 atribut awal yang digunakan, yaitu: berat jenis (<em>sg</em>), albumin (<em>al</em>), urea darah (<em>bu</em>), kreatinin serum (<em>sc</em>), sodium (<em>sod</em>), hemoglobin (<em>hemo</em>), sel darah merah (<em>rbc</em>), hipertensi (<em>htn</em>), diabetes mellitus (<em>dm</em>), dan nafsu makan (<em>appet</em>). Penggunaan atribut yang telah terseleksi tersebut, berhasil menekan biaya pemeriksaan hingga 73,36%. Selanjutnya dilakukan pendeteksian penyakit menggunakan Algoritme k-<em>Nearest Neighbor </em>menghasilkan nilai akurasi sebesar 99,25%, <em>sensitivity</em> sebesar 99,5%, dan <em>specificity</em> sebesar 98,745%.</p><p class="Abstrak"><em><strong>Abstract</strong></em></p><p class="Abstract"><em>Chronic kidney disease (CKD) is a health problem for people around the world with increasing incidence. Based on sources from BPJS Kesehatan, CKD care is the second largest ranking of financing after heart disease. CKD detection also requires many attributes, so it requires quite expensive costs. Create a system with web-based data mining stages that makes it easy to detect CKD. Allowing CKD to be prevented, addressed, and advised to get effective therapy is greater if acknowledged earlier. The process of this research uses work methods of Data Mining Knowledge Data Discovery (KDD). In the framework of the framework used, this system uses the Backward Elimination Algorithm to reduce the number of attributes used to reduce the type of inspection performed, and the k-Nearest Neighbor Algorithm as an algorithm to update disease. The best data mining modeling results from the system are made using Backward Elimination (α = 0.05) and kNN (k = 3) by calculating the increase in inspection costs and the highest sensitivity. System recommendations produce 10 attributes selected from the 24 initial attributes used, namely: specific gravity (sg), albumin (al), blood urea (bu), serum creatinine (sc), sodium (soil), hemoglobin (hemo), cell red blood (rbc), hypertension (htn), diabetes mellitus (dm), and appetite (appetite). The use of the selected attributes succeeded in achieving inspection costs of up to 73.36%. Furthermore, disease detection using the k-Nearest Neighbor Algorithm produces an accuracy value of 99.25%, sensitivity of 99.5%, and specificity of 98.745%.</em></p><p class="Abstrak"><em><strong><br /></strong></em></p>


2017 ◽  
Vol 4 (2) ◽  
pp. 134-142 ◽  
Author(s):  
Lucky Gagah Vedayoko ◽  
Endang Sugiharti ◽  
Much Aziz Muslim

Expert System is a computer system that has been entered the base of knowledge and set of rules to solve problems like an expert. One method in the expert system is Case Based Reasoning. To strengthen the retrieve stage of this method, the Nearest Neighbor algorithm is used. Bowel is one of the digestive organs susceptible to disease. The purpose of this study is to implement expert systems using Case Based Reasoning with Nearest Neighbor algorithm in diagnosing bowel disease and determine the accuracy of the system. Data used in this research are 60 data, obtained from medical record RSUD dr. Soetrasno Rembang. Variables used are general symptoms and types of diseases. The level of system accuracy resulting from scenario are 40 data as source case, and 20 data as target case that is equal to 95%.


Author(s):  
F. S. Ishaq ◽  
L. J. Muhammad ◽  
Yahaya B. Z ◽  
Abdurrahman Abubakar

Author(s):  
Oluwatosin Mayowa Alade ◽  
Olaperi Yeside Sowunmi ◽  
Sanjay Misra ◽  
Rytis Maskeliūnas ◽  
Robertas Damaševičius

Electronics ◽  
2020 ◽  
Vol 9 (12) ◽  
pp. 2117
Author(s):  
Xuesheng Peng ◽  
Ruizhi Chen ◽  
Kegen Yu ◽  
Feng Ye ◽  
Weixing Xue

The weighted K-nearest neighbor (WKNN) algorithm is the most commonly used algorithm for indoor localization. Traditional WKNN algorithms adopt received signal strength (RSS) spatial distance (usually Euclidean distance and Manhattan distance) to select reference points (RPs) for position determination. It may lead to inaccurate position estimation because the relationship of received signal strength and distance is exponential. To improve the position accuracy, this paper proposes an improved weighted K-nearest neighbor algorithm. The spatial distance and physical distance of RSS are used for RP selection, and a fusion weighted algorithm based on these two distances is used for position calculation. The experimental results demonstrate that the proposed algorithm outperforms traditional algorithms, such as K-nearest neighbor (KNN), Euclidean distance-based WKNN (E-WKNN), and physical distance-based WKNN (P-WKNN). Compared with the KNN, E-WKNN, and P-WKNN algorithms, the positioning accuracy of the proposed method is improved by about 29.4%, 23.5%, and 20.7%, respectively. Compared with some recently improved WKNN algorithms, our proposed algorithm can also obtain a better positioning performance.


Sign in / Sign up

Export Citation Format

Share Document