Maria Regina Justina E. Estuar

Validating the Voice of the Crowd During Disasters

Since the late 1990 s, the intensity of tropical cyclones have increased over time, causing massive flooding and landslides in thePhilippines. Nationwide Operational Assessment of Hazards or Project NOAH was put in place as a responsive program for disaster prevention and mitigation. Part of the solution was to set up nababaha.com(www.nababaha.com) and FloodPatrol which provided the public with a web and mobile phone based application for reporting flood height. This paper addresses the problem of providing an interactive and visual method of validating crowdsourced flood reports for the purpose of helping frontline responders and decision makers in disaster management. The approach involves finding the neighborhood of the crowdsourced flood report and weather station data based on their geospatial proximity and time record. A report is classified as correct if it falls within the obtained confidence interval of the crowdsourced flood report neighborhood. The neighborhood of crowdsourced flood reports are correlated with weather station data, which serves as the ground truth in the validation process. Use cases are presented to provide examples of automatic validation. The results of this study is beneficial to disaster management coordinators, first-line responders, government unit officials and citizens. The system provides an interactive approach in validating reports from the crowd, aside from providing an avenue to report flood events in an area. Overall, this contributes to the study of how crowdsourced reports are verified and validated.

Towards building a predictive model for remote river quality monitoring for mining sites

Most, if not all, mining sites in the Philippines are not equipped with expensive or modern monitoring tools to check for quality of soil, water and air elements which are relevant to ensure safety and wellness of miners. This study focused on the development of low cost mobile electronic sensors to monitor quality of water from rivers near mining sites. Low cost electronic sensors connected to a smart phone were developed to capture dissolved oxygen (DO2), pH, Turbidity, Temperature, and Salinity. The data for mercury (Hg) and arsenic (As) were obtained through AAS analyses to form baseline data for the model. Data was collected for over a period of one year, with site visits once every two months. A conditional inference tree (ctree) using recursive binary partitioning was used to generate the prediction model using 70 - 30 split on the training and test data set. The multi-feature model returns Good, Not Good or Unknown based on the scores of each element. The results showed a possible three feature model with significant results for site, salinity and pH balance.

Validating UI through UX in the Context of a Mobile - Web Crowdsourcing Disaster Management Application

A significant role of good user interface design is to enhance user experience. Keywords and icons are heavily used in mobile applications because of the limited screen size and small keypad interface. In creating a mobile based crowdsourcing application to motivate citizens to report disaster events and experiences, use of keywords and icons should be intuitive, have high recall and make reporting accurate. This paper discusses the user validation results of user interface design through user experience measured by intuitiveness, recall, and accuracy. For keywords, results showed that keywords were appropriate however there is a need to include selection of multiple keywords as well as language translated into native tongue. For icons, intuitiveness score were relatively high for most icons. Icons that had low recognition were because of multiple interpretations. Recall scores increased significantly after training. Enhancing user experience using simulation scenarios increased responsiveness but decreased correct submissions. This study shows that engagement increases if simulations depict real scenarios making the system relevant to its intended users.

Microscopic Fusarium Detection and Verification with Convolutional Neural Networks

Advances in computer vision, specifically on deep convolutional neural networks, have achieved state of the art results in multiple computer vision tasks. These networks have enabled the rapid de-tection of Fusarium oxysporum, the main cause of the Fusarium Wilt disease plaguing banana Cavendish plantations. The paper focuses on the use of convolutional neural networks and deep learn-ing by training and fine-tuning a MobileNet-based deep learning model for Fusarium detection in microscopy images. Multiple im-age augmentation techniques have been used to induce feature invariance in the model to control for the unique characteristics of Fusarium. Analysis of the behavior of the trained model using Locally Interpretable Model-agnostic Explanations, or LIME, has been performed to verify correct behavior. Results with MobileNet and LIME have shown that two out of four models have been able to specifically discriminate Fusarium from other present artifacts, such as soil particles.

Estimating parameters for a dynamical dengue model using genetic algorithms

Dynamical models are a mathematical framework for understanding the spread of a disease using various epidemiological parameters. However, in data-scarce regions like the Philippines, local estimates of epidemiological parameters are difficult to obtain because methods to obtain these values are costly or inaccessible. In this paper, we employ genetic algorithms trained with novel fitness functions as a low-cost, data-driven method to estimate parameters for dengue incidence in the Western Visayas Region of the Philippines (2011-2016). Initial results show good ht between monthly historical values and model outputs using parameter estimates, with a best Pearson correlation of 0.86 and normalized error of 0.65 over the selected 72-month period. Furthermore, we demonstrate a quality assessment procedure for selecting biologically feasible and numerically stable parameter estimates. Implications of our findings are discussed in both epidemiological and computational contexts, highlighting their application in FASSSTER, an integrated syndromic surveillance system for infectious diseases in the Philippines.

Agent-Based Modeling Approach in Understanding Behavior During Disasters: Measuring Response and Rescue in eBayanihan Disaster Management Platform

Development of a disaster management system is as complex as the environment it mimics. In 2015, the eBayanihan disaster management platform was launched in Metro Manila, Philippines. It is designed to be an integrated multidimensional and multi-platform system that can be used in managing the flow of information during disaster events. Since its development, usage of the system varies depending on the agent who uses the system and which area is affected by what type of disaster. As a complex problem, behavior of disaster agents, such as official responders, volunteers, regular citizens, is best understood if the system can capture, model, and visualize behavior over time. This study presents the development and implementation of an agent-based approach in understanding disaster response and rescue by automatically capturing agent behavior in the eBayanihan Disaster Management Platform. All user activities are logged and converted into behavior matrices that can be saved and imported into the Organizational Risk Analyzer (ORA) tool. ORA is used to generate the agent-based model which can be viewed in the eBayanihan platform. Actual behavior (ABehM) is compared against perceived (PBM) and expected behavior (EBM) during rescue and response. Results show that EBM networks are fully connected while PBM during rescue and response are granular and vast. Both however show centrality at the provincial and municipal level. ABehM on the other hand shows concentration only at the municipal level with more interactions with ordinary volunteers and citizens.

Nepal: The Role of Nurses After Nepal Earthquake 2015

The role of a professional nursing organization, that is, the Nursing Association of Nepal (NAN), in disaster management after the devastating earthquake in April 25, 2015, shows how nurses supported a technological innovation in managing health information in the country during and after the disaster. Nurses were trained on rapid health assessment and use of electronic health records and referral system and a surveillance system to help in gathering information on the ground to inform health interventions.

Comparative analysis of tree classification models for detecting fusarium oxysporum f. sp cubense (TR4) based on multi soil sensor parameters

Use of wireless sensor networks and smartphone integration design to monitor environmental parameters surrounding plantations is made possible because of readily available and affordable sensors. Providing low cost monitoring devices would be beneficial, especially to small farm owners, in a developing country like the Philippines, where agriculture covers a significant amount of the labor market. This study discusses the integration of wireless soil sensor devices and smartphones to create an application that will use multidimensional analysis to detect the presence or absence of plant disease. Specifically, soil sensors are designed to collect soil quality parameters in a sink node from which the smartphone collects data from via Bluetooth. Given these, there is a need to develop a classification model on the mobile phone that will report infection status of a soil. Though tree classification is the most appropriate approach for continuous parameter-based datasets, there is a need to determine whether tree models will result to coherent results or not. Soil sensor data that resides on the phone is modeled using several variations of decision tree, namely: decision tree (DT), best-fit (BF) decision tree, functional tree (FT), Naive Bayes (NB) decision tree, J48, J48graft and LAD tree, where decision tree approaches the problem by considering all sensor nodes as one. Results show that there are significant differences among soil sensor parameters indicating that there are variances in scores between the infected and uninfected sites. Furthermore, analysis of variance in accuracy, recall, precision and F1 measure scores from tree classification models homogeneity among NBTree, J48graft and J48 tree classification models.

Streamlining machine learning in mobile devices for remote sensing

Mobile devices have been at the forefront of Intelligent Farming because of its ubiquitous nature. Applications on precision farming have been developed on smartphones to allow small farms to monitor environmental parameters surrounding crops. Mobile devices are used for most of these applications, collecting data to be sent to the cloud for storage, analysis, modeling and visualization. However, with the issue of weak and intermittent connectivity in geographically challenged areas of the Philippines, the solution is to provide analysis on the phone itself. Given this, the farmer gets a real time response after data submission. Though Machine Learning is promising, hardware constraints in mobile devices limit the computational capabilities, making model development on the phone restricted and challenging. This study discusses the development of a Machine Learning based mobile application using OpenCV libraries. The objective is to enable the detection of Fusarium oxysporum cubense (Foc) in juvenile and asymptomatic bananas using images of plant parts and microscopic samples as input. Image datasets of attached, unattached, dorsal, and ventral views of leaves were acquired through sampling protocols. Images of raw and stained specimens from soil surrounding the plant, and sap from the plant resulted to stained and unstained samples respectively. Segmentation and feature extraction techniques were applied to all images. Initial findings show no significant differences among the different feature extraction techniques. For differentiating infected from non-infected leaves, KNN yields highest average accuracy, as opposed to Naive Bayes and SVM. For microscopic images using MSER feature extraction, KNN has been tested as having a better accuracy than SVM or Naive-Bayes.

Development of an asynchronous communication channel between wireless sensor nodes, smartphone devices, and web applications using RESTful Web Services for intelligent farming

Environment and agriculture related applications have been gaining ground for the past several years and have been the context for researches in ubiquitous and pervasive computing. This study is a part of a bigger study that uses artificial intelligence in developing models to detect, monitor, and forecast the spread of Fusarium oxysporum cubense TR4 (FOC TR4) on Cavendish bananas cultivated in the Philippines. To implement an Intelligent Farming system, 1) wireless sensor nodes (WSNs) are deployed in Philippine banana plantations to collect soil parameter data that is considered to affect the health of Cavendish bananas, 2) a custom built smartphone application is used for collecting, storing, and transmitting soil data, plant images and plant status data to a cloud storage, and 3) a custom built web application is used to load and display results of physico-chemical analysis of soil, analysis of data models, and geographic locations of plants being monitored. This study discusses the issues, considerations, and solutions implemented in the development of an asynchronous communication channel to ensure that all data collected by WSNs and smartphone applications are transmitted with a high degree of accuracy and reliability. From a design standpoint: standard API documentation on usage of data type is required to avoid inconsistencies in parameter passing. From a technical standpoint, there is a need to include error-handling mechanisms especially for delays in transmission of data as well as generalize method of parsing thru multidimensional array of data. Strategies are presented in the paper.