Spatiotemporal retrieval and feature analysis of air pollution episodes.

Air pollution has inevitably come along with the economic development of human society. How to balance economic growth with a sustainable environment has been a global concern. The ambient PM2.5 (particulate matter with aerodynamic diameter ≤ 2.5 µm) is particularly life-threatening because these tiny aerosols could be inhaled into the human respiration system and cause millions of premature deaths every year. The focus of most relevant research has been placed on apportionment of pollutants and the forecast of PM2.5 concentration measures. However, the spatiotemporal variations of pollution regions and their relationships to local factors are not much contemplated in the literature. These local factors include, at least, land terrain, meteorological conditions and anthropogenic activities. In this paper, we propose an interactive analysis platform for spatiotemporal retrieval and feature analysis of air pollution episodes. A domain expert can interact with the platform by specifying the episode analysis intention considering various local factors to reach the analysis goals. The analysis platform consists of two main components. The first component offers a query-by-sketch function where the domain expert can search similar pollution episodes by sketching the spatial relationship between the pollution regions and the land objects. The second component helps the domain expert choose a retrieved episode to conduct spatiotemporal feature analysis in a time span. The integrated platform automatically searches the episodes most resembling the domain expert's original sketch and detects when and where the episode emerges and diminishes. These functions are helpful for domain experts to infer insights into how local factors result in particular pollution episodes.

Ensemble learning of model hyperparameters and spatiotemporal data for calibration of low-cost PM2.5 sensors.

he PM2.5 air quality index (AQI) measurements from government-built supersites are accurate but cannot provide a dense coverage of monitoring areas. Low-cost PM2.5 sensors can be used to deploy a fine-grained internet-of-things (IoT) as a complement to government facilities. Calibration of low-cost sensors by reference to high-accuracy supersites is thus essential. Moreover, the imputation for missing-value in training data may affect the calibration result, the best performance of calibration model requires hyperparameter optimization, and the affecting factors of PM2.5 concentrations such as climate, geographical landscapes and anthropogenic activities are uncertain in spatial and temporal dimensions. In this paper, an ensemble learning for imputation method selection, calibration model hyperparameterization, and spatiotemporal training data composition is proposed. Three government supersites are chosen in central Taiwan for the deployment of low-cost sensors and hourly PM2.5 measurements are collected for 60 days for conducting experiments. Three optimizers, Sobol sequence, Nelder and Meads, and particle swarm optimization (PSO), are compared for evaluating their performances with various versions of ensembles. The best calibration results are obtained by using PSO, and the improvement ratios with respect to R2, RMSE, and NME, are 4.92%, 52.96%, and 56.85%, respectively.

Integrating relevance feedback techniques for image retrieval using reinforcement learning.

Relevance feedback (RF) is an interactive process which refines the retrievals to a particular query by utilizing the user's feedback on previously retrieved results. Most researchers strive to develop new RF techniques and ignore the advantages of existing ones. In this paper, we propose an image relevance reinforcement learning (IRRL) model for integrating existing RF techniques in a content-based image retrieval system. Various integration schemes are presented and a long-term shared memory is used to exploit the retrieval experience from multiple users. Also, a concept digesting method is proposed to reduce the complexity of storage demand. The experimental results manifest that the integration of multiple RF approaches gives better retrieval performance than using one RF technique alone, and that the sharing of relevance knowledge between multiple query sessions significantly improves the performance. Further, the storage demand is significantly reduced by the concept digesting technique. This shows the scalability of the proposed model with the increasing-size of database.