Extreme Learning Machine Prediction Model on Airbnb Base Price

FIKRI NURQAHHARI P, Dr. Agus Sihabuddin, S.Si., M.Kom.

2020 | Skripsi | S1 ILMU KOMPUTER

This paper describes how Extreme Learning Machine is implemented as a prediction model to determine the base price of Airbnb properties. Extreme Learning Machine offers many advantages, such as fast learning speed, good generalization performance, and high prediction accuracy. The steps in the extreme learning machine method, in general, are set hidden neuron number, randomly assign input weight & hidden layer biases, calculate the output layer, the whole learning process completed through one mathematical change without iteration. The performance of the model will be measured using mean absolute percentage error (MAPE), mean squared error (MSE), and root mean squared error (RMSE). The ELM model will be tested with the Airbnb dataset in the London area taken from the Insideairbnb.com site with 79671 numbers of data with 21 features chosen. MAPE, MSE, RMSE are chosen for the evaluation method. From the experiments, the model show as the number of neurons increases, the link between the input and output layers would consequently increase. This leads to a better quality of learning produced from and the model generates MAPE value of 3.06% MSE value of 0. 038, and RMSE value of 0.239.

This paper describes how Extreme Learning Machine is implemented as a prediction model to determine the base price of Airbnb properties. Extreme Learning Machine offers many advantages, such as fast learning speed, good generalization performance, and high prediction accuracy. The steps in the extreme learning machine method, in general, are set hidden neuron number, randomly assign input weight & hidden layer biases, calculate the output layer, the whole learning process completed through one mathematical change without iteration. The performance of the model will be measured using mean absolute percentage error (MAPE), mean squared error (MSE), and root mean squared error (RMSE). The ELM model will be tested with the Airbnb dataset in the London area taken from the Insideairbnb.com site with 79671 numbers of data with 21 features chosen. MAPE, MSE, RMSE are chosen for the evaluation method. From the experiments, the model show as the number of neurons increases, the link between the input and output layers would consequently increase. This leads to a better quality of learning produced from and the model generates MAPE value of 3.06% MSE value of 0. 038, and RMSE value of 0.239.

Kata Kunci : Airbnb, Extreme Learning Machine, Price Prediction