Predicting Catastrophic Events Using Machine Learning Models for Natural Language Processing

In recent years, there has been widespread improvement in communication technologies. Social media applications like Twitter have made it much easier for people to send and receive information. A direct application of this can be seen in the cases of disaster prediction and crisis. With people being able to share their observations, they can help spread the message of caution. However, the identification of warnings and analyzing the seriousness of text is not an easy task. Natural language processing (NLP) is one way that can be used to analyze various tweets for the same. Over the years, various NLP models have been developed that are capable of providing high accuracy when it comes to data prediction. In the chapter, the authors will analyze various NLP models like logistic regression, naive bayes, XGBoost, LSTM, and word embedding technologies like GloVe and transformer encoder like BERT for the purpose of predicting disaster warnings from the scrapped tweets. The authors focus on finding the best disaster prediction model that can help in warning people and the government.

Rainstorm Disaster Prediction Based on Weather Research and Forecast (WRF): A Case of Nanjing, China

Rainstorm disaster causes great damage to human lives, environment and economies. Many environmental catastrophes happened every summer in southern China as result of the rainstorm. Therefore, heavy rain prediction remains the main focus of many scholars’ attention. However, the weather forecast is inaccurate and not prompt enough, causing casualties and financial losses. Weather Research Forecast (WRF) is an effective method and is utilized in this study to predict the precipitable water vapor (PWV) in meso-and small-scale in Nanjing. Rain is formed because of the PWV in the atmosphere, and therefore precipitation could be predicted according to the PWV. A method is proposed that the amount of rainstorm precipitation could be predicted based on the PWV, which can be simulated by WRF. The experimental results are consistent with the actual rainstorm situation. It demonstrates that promising measures based on the reliable WRF model could be taken to reduce the impending disasters.

A Sentiment-Aware Contextual Model for Real-Time Disaster Prediction Using Twitter Data

The massive amount of data generated by social media present a unique opportunity for disaster analysis. As a leading social platform, Twitter generates over 500 million Tweets each day. Due to its real-time characteristic, more agencies employ Twitter to track disaster events to make a speedy rescue plan. However, it is challenging to build an accurate predictive model to identify disaster Tweets, which may lack sufficient context due to the length limit. In addition, disaster Tweets and regular ones can be hard to distinguish because of word ambiguity. In this paper, we propose a sentiment-aware contextual model named SentiBERT-BiLSTM-CNN for disaster detection using Tweets. The proposed learning pipeline consists of SentiBERT that can generate sentimental contextual embeddings from a Tweet, a Bidirectional long short-term memory (BiLSTM) layer with attention, and a 1D convolutional layer for local feature extraction. We conduct extensive experiments to validate certain design choices of the model and compare our model with its peers. Results show that the proposed SentiBERT-BiLSTM-CNN demonstrates superior performance in the F1 score, making it a competitive model in Tweets-based disaster prediction.

A Study on Establishment of Drought Index for Agricultural Disaster Prediction

This study aims to establish a drought index for disaster prediction in Gyeongsangnam-do, where the most agricultural drought damage occurred from 1965 to 2018. The drought index was analyzed for each duration (3, 6, 9, 12 months) targeting the SPI. Damage characteristics of the duration of agricultural drought were calculated. SPI for each duration of agricultural drought damage period in Gyeongsangnam-do was at least -2.0 or less, and the maximum was -1.0 or more, and weak and moderate drought were analyzed. However, due to the heavy rain effect during the rainy season, the average SPI12 was -1.06, and the impact of agricultural drought was negligible. It was analyzed that the correlation between the damage period of agricultural drought and the SPI by duration was high. However, there is not much difference in SPI for each duration to determine the occurrence of damage. In this study, the criterion for disaster prediction of agricultural drought was calculated as representative drought index by year as the minimum drought index of SPI for each duration of damage occurrence period of past agricultural drought. The Standard of drought index for disaster prediction was set to -1.64, the average of the SPI for each duration of year in which damage occurred in the past.