Using self-organizing maps as unsupervised learning models for meteorological data mining

With Internet of Things (IoT) being prevalently adopted in recent years, traditional machine learning and data mining methods can hardly be competent to deal with the complex big data problems if applied alone. However, hybridizing those who have complementary advantages could achieve optimized practical solutions. This work discusses how to solve multivariate regression problems and extract intrinsic knowledge by hybridizing Self-Organizing Maps (SOM) and Regression Trees. A dual-layer SOM map is developed in which the first layer accomplishes unsupervised learning and then regression tree layer performs supervised learning in the second layer to get predictions and extract knowledge. In this framework, SOM neurons serve as kernels with similar training samples mapped so that regression tree could achieve regression locally. In this way, the difficulties of applying and visualizing local regression on high dimensional data are overcome. Further, we provide an automated growing mechanism based on a few stop criteria without adding new parameters. A case study of solving Electrical Vehicle (EV) range anxiety problem is presented and it demonstrates that our proposed hybrid model is quantitatively precise and interpretive. key words: Multivariate Regression, Big Data, Machine Learning, Data Mining, Self-Organizing Maps (SOM), Regression Tree, Electrical Vehicle (EV), Range Estimation, Internet of Things (IoT)

Download Full-text

Analisis Cluster Data Interkomparasi Anak Timbangan dengan Algoritma Self Organizing Maps

Jurnal Teknik Informatika dan Sistem Informasi ◽

10.28932/jutisi.v7i2.3698 ◽

2021 ◽

Vol 7 (2) ◽

Author(s):

Arif Fajar Solikin ◽

Kusrini Kusrini ◽

Ferry Wahyu Wibowo

Keyword(s):

Data Mining ◽

Statistical Test ◽

Optimum Number ◽

Data Normalization ◽

Self Organizing Map ◽

Self Organizing Maps ◽

Number Of Clusters ◽

Map Algorithm ◽

Cluster Data ◽

Self Organizing

Intercomparison was conducted to determine the ability and the performance of the laboratory. Intercomparison results are usually expressed in the range of En ratio values (En ?|1|) which express the equivalence of one laboratory with other laboratories. If the laboratory is declared unequal, then it needs to identify the source of the problem by itself. To make it easier, it can be done by Clustering which is one of the data mining techniques. Clustering is done by applying a self organizing map algorithm on the KNIME (Konstanz Information Miner) analytic tools. Several experiments were carried out with different layer size and data normalization status from one experiment to another experiment. The results were analyzed through pseudo F statistical test and icdrate test. The largest pseudo F statistic value was obtained from the 8th experiment (setting the layer size 2x2 without data normalization) with a pseudo F statistic value of 167.53 for 1kg artifacts and a Pseudo F statistic value of 104.86 for 200 g artifacts where the optimum number of clusters are 4. The smallest icdrate value was obtained from the 5th experiment (setting the 2x3 layer size without data normalization) with an icdrate value of 0.0713 for 1kg artifacts and icdrate value of 0.2889 for 200g artifacts with the best number of clusters being 6. From 12 laboratories can be grouped into 6 groups where each group has the same identification. There are groups 1, 3 and 6 have 1 member, while groups 2, 4 and 5 have 3 members.

Download Full-text

Ketepatan Klasifikasi Penerima Beasiswa STMIK STIKOM Bali dengan Hybrid Self Organizing Maps dan Algoritma K-Mean

Jurnal Varian ◽

10.30812/varian.v2i1.316 ◽

2018 ◽

Vol 2 (1) ◽

pp. 1-7 ◽

Cited By ~ 1

Author(s):

Ni Putu Nanik Hendayanti ◽

Gusti Ayu Made Arna Putri ◽

Maulida Nurhidayati

Keyword(s):

Neural Network ◽

Data Mining ◽

Artificial Neural Network ◽

Self Organizing Maps ◽

Cluster Group ◽

Artificial Neural ◽

Clustering Data ◽

Self Organizing

Data Mining adalah penemuan informasi baru dengan mencari pola atau aturan tertentu dari sejumlah data yang sangat besar. Salah satu teknik yang dikenal dalam Data Mining yaitu clustering. Pengertian clustering dalam Data Mining adalah pengelompokan sejumlah data atau objek ke dalam cluster (group) sehingga setiap di lama cluster tersebut akan berisi data yang semirip mungkin dan berbeda dengan objek dalam cluster yang lain. Salah satu metode klasifiaksi atau clustering adalah Self Organizing Maps (SOM). SOM merupakan metode artificial neural network yang digunakan untuk mengelompokkan (clustering) data berdasarkan karakteristik/fitur-fitur data. Metode pengelompokan yang menggunakan konsep jarak dan memiliki karakteristik yang hampir sama dengan SOM yaitu metode K-means. Penelitian ini bertujuan untuk mengembangkan suatu metode yang merupakan hybrid dari SOM dan K-means yang digunakan untuk menentukan ketepatan suatu klasifikasi. Sebelum diujikan pada data asli, metode hybrid SOM dan K-Means diujikan lebih dulu pada data benchmark sehingga dapat diketahui berapa persen ketepan yang dihasilkan. Kemudian dilanjutkan dengan penerapan metode hybrid SOM dan K-means pada data penerimaan beasiswa di STMIK STIKOM Bali. Penelitian ini bertujuan untuk menentukan ketepatan klasifikasi penerima beasiswa STMIK STIKOM Bali dengan metode hybrid SOM dan K-means. Hasil penelitian menunjukkan bahwa metode Kmeans dan SOM memberikan hasil yang sama yang akibatnya metode SOM-Kmeans juga memberikan hasil yang sama. Alasannya, metode SOM-Kmeans menggunakan nilai centroid dari hasil SOM, dan hasil yang diperoleh pada metode Kmean memiliki hasil yang sama dengan SOM akibatnya metode SOM-Kmeans menghasilkan hasil yang sama dengan kedua metode sebelumnya.

Download Full-text