Klasifikasi Aktivitas Manusia Menggunakan Algoritme Computed Input Weight Extreme Learning Machine dengan Reduksi Dimensi Principal Component Analysis

Penulis

  • M. Sofyan Irwanto Universitas Brawijaya, Malang
  • Fitra A. Bachtiar Universitas Brawijaya, Malang
  • Novanto Yudistira Universitas Brawijaya, Malang

DOI:

https://doi.org/10.25126/jtiik.2022965504

Abstrak

Salah satu bidang penelitian yang sangat penting yaitu pengenalan aktivitas manusia secara otomatis dikarenakan potensi penerapannya di berbagai bidang lain seperti pengawasan, lingkungan cerdas, maupun kesehatan. Dari berbagai pendekatan yang pernah dilakukan untuk mengenali aktivitas manusia, teknik berbasis sensor diketahui lebih unggul daripada teknik lain seperti teknik berbasis visi komputer. Teknik berbasis sensor juga dapat dilakukan menggunakan ponsel cerdas, namun penggunaan ponsel cerdas memiliki kekurangan dalam melakukan komputasi algoritme yang kompleks. Apalagi, data hasil ekstraksi fitur dari sinyal yang ditangkap oleh sensor memiliki dimensi yang tinggi. Sehingga, diperlukan sebuah metode untuk mengurangi dimensi fitur dari data, serta melakukan klasifikasi terhadap data tersebut dengan cepat dan tepat. Salah satu metode yang dapat digunakan untuk mengurangi dimensi fitur dari sebuah data adalah Principal Component Analysis (PCA), dan salah satu metode klasifikasi yang dapat digunakan adalah Computed Input Weight Extreme Learning Machine (CIW-ELM). Oleh karena itu, penelitian ini akan menggunakan kedua metode tersebut untuk melakukan klasifikasi pada aktivitas sederhana seperti berjalan, menaiki tangga, menuruni tangga, duduk, berdiri, dan berbaring. Pada penelitian ini juga dilakukan pemilihan hyperparameter terbaik pada masing-masing metode menggunakan metode Grid Search Cross Validation. Hyperparameter terbaik yang didapatkan untuk algoritme PCA adalah dengan nilai k = 207, serta untuk algoritme CIW-ELM dengan jumlah hidden neuron = 600 dan fungsi aktivasi sigmoid. Hasil akurasi yang didapatkan pada penelitian ini adalah 0,957 dan rata-rata f-measure sebesar 0,958 dengan waktu pelatihan selama 0,57 detik.

 

Abstract

 

One of the most important research area is automatic human activity recognition due to its potential application in various other fields such as surveillance, smart environment, and healthcare. Based on various approaches that have been used to recognize human activity, sensor-based techniques are known to be superior to other techniques such as computer vision-based techniques. Sensor-based technique can also be performed using smartphones, but smartphone has disadvantages in performing complex alghorithmic computation. Moreover, feature extraction of the data from the signal captured by the sensor has high dimensions. So, we need a methods to reduce the feature dimensions of the data, and classify the data quickly and accurately. One of the method that can be used to reduce the feature dimensions of data is Principal Component Analysis (PCA), and one of the classification methods that can be used is Computed Input Weight Extreme Learning Machine (CIW-ELM). Therefore, this study will use both methods to classify simple activities such as walking, walking upstairs, walking downstairs, sitting, standing, and laying. In this study, the selection of the best hypeparameter for each method was also carried out using Grid Search Cross Validation. The best hyperparameter obtained for the PCA algorithm is with a value of k = 207, and for the CIW-ELM algorithm with the number of hidden neurons = 600 and the sigmoid activation function. The accuracy results obtained in this study were 0,957 and the f-measure average were 0,958 with a training time of 0,57 seconds.

Downloads

Download data is not yet available.

Referensi

AGGARWAL, J. K., & RYOO, M. S., 2011. Human activity analysis: A review. ACM Computing Surveys, 43(3).

AHMAD, I., BASHERI, M., IQBAL, M. J., & RAHIM, A. 2018. Performance Comparison of Support Vector Machine, Random Forest, and Extreme Learning Machine for Intrusion Detection. IEEE Access, 6(c), 33789–33795.https://doi.org/10.1109/ACCESS.2018.2841987

ANGUITA, D., GHIO, A., ONETO, L., PARRA, X., & REYES-ORTIZ, J. L., 2013. A Public Domain Dataset for Human Activity Recognition Using Smartphone. ESANN 2013 Proceedings, European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. Bruges (Belgium), 24-26 April 2013, I6doc.Com, April, 437–442.

BEVILACQUA, A., MACDONALD, K., RANGAREJ, A., WIDJAYA, V., CAULFIELD, B., & KECHADI, T., 2019. Human activity recognition with convolutional neural networks. In arXiv (Vol. 1). Springer International Publishing.

DOEWES, A., SWASONO, S. E., & HARJITO, B., 2017. Feature selection on Human Activity Recognition dataset using Minimum Redundancy Maximum Relevance. 2017 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2017, 1, 171–172.

GARCIA-GONZALEZ, D., RIVERO, D., FERNANDEZ-BLANCO, E., & LUACES, M. R., 2020. A public domain dataset for real-life human activity recognition using smartphone sensors. Sensors (Switzerland), 20(8).

HERNANDEZ, F., SUAREZ, L. F., VILLAMIZAR, J., & ALTUVE, M., 2019. Human Activity Recognition on Smartphones Using a Bidirectional LSTM Network.

HUANG, G. BIN, ZHU, Q. Y., & SIEW, C. K., 2004. Extreme learning machine: A new learning scheme of feedforward neural networks. IEEE International Conference on Neural Networks - Conference Proceedings, 2(February 2014), 985–990.

JOLLIFE, I. T., & CADIMA, J., 2016. Principal component analysis: A review and recent developments. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 374(2065).

JUNITA, V., & BACHTIAR, F. A., 2019. Klasifikasi Aktivitas Manusia menggunakan Algoritme Decision Tree C4 . 5 dan Information Gain untuk Seleksi Fitur. 3(10), 9426–9433.

KHAN, A. M., LEE, Y. K., LEE, S. Y., & KIM, T. S., 2010. A triaxial accelerometer-based physical-activity recognition via augmented-signal features and a hierarchical recognizer. IEEE Transactions on Information Technology in Biomedicine, 14(5), 1166–1172.

MICUCCI, D., MOBILIO, M., & NAPOLETANO, P., 2017. UniMiB SHAR: A dataset for human activity recognition using acceleration data from smartphones. Applied Sciences (Switzerland), 7(10).

POTDAR, K., S., T., & D., C., 2017. A Comparative Study of Categorical Variable Encoding Techniques for Neural Network Classifiers. International Journal of Computer Applications, 175(4), 7–9

REISS, A., HENDEBY, G., & STRICKER, D., 2013. A competitive approach for human activity recognition on smartphones. ESANN 2013 Proceedings, 21st European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, April, 455–460.

SU, X., TONG, H., & JI, P., 2014. Activity recognition with smartphone sensors. Tsinghua Science and Technology, 19(3), 235–249.

TAPSON, J., DE CHAZAL, P., & VAN SCHAIK, A., 2015. Explicit Computation of Input Weights in Extreme Learning Machines. June, 41–49.

WANG, A., CHEN, G., YANG, J., ZHAO, S., & CHANG, C., 2016. A Comparative Study on Human Activity Recognition Using Inertial Sensors in a Smartphone.

WANG, B., CHEN, Q., WANG, Z., & HU, Y., 2019. The Rearch on Improved LVQ Neural Network Method. 2019 3rd International Conference on Circuits, System and Simulation, ICCSS 2019, 206–209.

ZHANG, M., & SAWCHUK, A. A., 2012. A feature selection-based framework for human activity recognition using wearable multimodal sensors. BODYNETS 2011 - 6th International ICST Conference on Body Area Networks, January 2011, 92–98. https://doi.org/10.4108/icst.bodynets.2011.247018

Diterbitkan

22-12-2022

Terbitan

Bagian

Ilmu Komputer

Cara Mengutip

Klasifikasi Aktivitas Manusia Menggunakan Algoritme Computed Input Weight Extreme Learning Machine dengan Reduksi Dimensi Principal Component Analysis. (2022). Jurnal Teknologi Informasi Dan Ilmu Komputer, 9(6), 1195-1202. https://doi.org/10.25126/jtiik.2022965504