A Study of Pattern Prediction in the Monitoring Data of Earthen Ruins with the Internet of Things

Abstract

An understanding of the changes of the rammed earth temperature of earthen ruins is important for protection of such ruins. To predict the rammed earth temperature pattern using the air temperature pattern of the monitoring data of earthen ruins, a pattern prediction method based on interesting pattern mining and correlation, called PPER, is proposed in this paper. PPER first finds the interesting patterns in the air temperature sequence and the rammed earth temperature sequence. To reduce the processing time, two pruning rules and a new data structure based on an R-tree are also proposed. Correlation rules between the air temperature patterns and the rammed earth temperature patterns are then mined. The correlation rules are merged into predictive rules for the rammed earth temperature pattern. Experiments were conducted to show the accuracy of the presented method and the power of the pruning rules. Moreover, the Ming Dynasty Great Wall dataset was used to examine the algorithm, and six predictive rules from the air temperature to rammed earth temperature based on the interesting patterns were obtained, with the average hit rate reaching 89.8%. The PPER and predictive rules will be useful for rammed earth temperature prediction in protection of earthen ruins.

Composite Nano-Calcium-Based Material for Reinforcement and Protection of Earthen Ruins in Humid Environment

“…Experiments were carried out to prove the accuracy of the proposed method and the power of the pruning rules. In addition, the algorithm was tested using the Great Wall dataset of the Ming Dynasty, and 6 prediction rules from temperature to rammed Earth temperature were obtained according to interesting models, and the average hit rate reached 89.8% [1]. Dong et al article proposes a joint method to quantitatively analyze the correlation between the monitored environmental factors and the degradation of soil sites in humid areas caused by water saturation.…”

Improving Automatic Fish Freshness Detection using TCS 230 and Support Vector Machine

“…Penelitian lainnya juga menggunakan sensor kristal kuarsa mikro (QCM) dan algoritma Support Vector Machine (SVM) untuk memprediksi kesegaran ikan dengan mendeteksi aroma yang dihasilkan dari ikan segar dan tidak segar. Dari data sensor QCM yang diolah menggunakan SVM, sistem yang dikembangkan mampu mengklasifikasikan tingkat kesegaran ikan dengan akurasi sekitar 92%, menunjukkan potensi teknologi QCM dan SVM sebagai alat efektif untuk memprediksi kesegaran ikan dalam skala industri [7]. Hasil penelitian sebelumnya menunjukkan bahwa sistem yang dikembangkan dapat mendeteksi kesegaran ikan dengan tingkat akurasi yang tinggi.…”