Search

Air pollution is a product of the development of civilization. With the development of society, the diseases and injuries caused by PM2.5 (fine particulate matter) to the human body have become an important issue that countries all over the world want to solve. At present, most air quality index AQI monitoring uses machine learning and deep learning such as LSTM to predict the trend of air pollution, but this method causes the prediction model to always be one hour away from the actual value, which makes it impossible to judge some abnormal turning changes in time. Therefore, this article uses the following methods: (1) Use the deep learning method CNN-GRU classification model for training, use the data of the surrounding micro-sites in the previous hour as input, and the change trend of PM2.5 as output. (2) Using the characteristics of the micro-station changes, and try to select the range of the micro-stations according to the wind direction. We expect to find the characteristics and relationships between micro-stations and the central EPA observation station. We obtain the following change trend of PM2.5 within one hour, and then use the change trend to modify the predicted value of the original prediction model. Although the desired result has not yet been achieved, in the future, the use of wind speed to select the range of the micro-station may be considered, or other methods may be used as the direction of follow-up research.

The assessment of mental state is an important issue for diagnosing whether people getting depressed or not. The general assessment tool is to use the Hamilton Depression Rating Scale (HAM-D). In general, psychiatrist utilizes Monthly Mood Chart (MMC) to collect and infer the mental state during the treatment for assessing whether the symptoms relieved or not. But, the MMC is a kind of psychology data. Recently, there is research has developed to evaluate individual emotional and mental state in the way of fusion of psychology and physiology. Therefore, how to effectively and conveniently collect patient's psychological and physiological data to integrate them together has become an important issue. In this paper, we design and implement a platform for collecting data and inferring mental state; the data includes patients' physiological and psychological data, such as Electroencephalography (EEG), Heart Rate Variability (HRV), and MMC. The platform can automatically infer emotional state and provide it to psychiatrists for assessment. This platform is cooperating with psychiatrists in two hospitals and would like to obtain approval. It can be used for the diagnosis of future mental illnesses.

In recent years, the downside risk measure is considered to be more in line with investors’ attitudes towards risk. The purpose of this paper is to establish a diversified portfolio selection model under the downside risk framework. Firstly, a fuzzy mean semi-absolute deviation-semi-variance-proportional entropy portfolio selection model with transaction costs is proposed, based on Markowitz’s mean-variance model, in which semi-absolute deviation and semi-variance are taken as the risk measure. That is to say, the possibilistic mean of the investment rate of return is used to quantify the return of the portfolio, and the possibilistic semi-absolute deviation and possibilistic semi-variance are used to measure investment risk. The shortcomings of semi-absolute deviation or semi-variance as a risk measure individually can be overcome when the semi-absolute deviation is combined with semi-variance as a dual risk measure. Meanwhile, a proportional entropy function based on the Minkowski measure is used to construct the proposed model to diversify the portfolio. Secondly, a new multi-objective bat algorithm is designed to solve the model, and the experiments were conducted by using data from the Shanghai Stock Exchange to prove the effectiveness of the model.

Nearest Level Modulation and capacitor voltage sorting algorithm are the most commonly-used modulation strategies for MMC in engineering. Unlike traditional carrier-based modulation methods, the Nearest Level Modulation and capacitor voltage sorting algorithm may make the switching sequences irregular, bringing difficulties to the analytic modeling of EMf sources wave forms. Besides, the extremely short simulation step size and long triansient state lead to huge time consumption on time domain EMf predictions. To solve these problems, a fast modeling method for EMf sources in MMC based on Nearest Level Modulation and steady-state capacitor voltage sorting algorithm is proposed for frequency domain EMf predictions. Time and frequency domain results show great agreement with Matlab Simulink simulation. The proposed method can be a reference for further research on the EMf predictions of MMC.

AC/DC hybrid distribution network has attracted considerable research interest in recent years, and the coordinated control of multiple VSCs is a key issue. This paper proposes a novel adaptive droop control strategy based on DC voltage deviation and power imbalance, which ensures the DC voltage brief and small over-limit during transient process and no steady-state deviation, with the power margin of the system is amplified as well. Case study and simulation are performed based on the MATLAB model of a three-terminal AC/DC hybrid distribution network system to demonstrate the effectiveness and superiority of the proposed adaptive droop control.

This paper proposes a novel voltage balancing method for Modular Multilevel Converter (MMC) under Phase- Shifted Carrier Pulsewidth Modulation (CPS-PWM). For the existing individual voltage balancing methods, an unsolved problem is that the active power transmission will greatly affect the balancing speed of MMC system. By utilizing the output current amplitude information, the effect of working condition on voltage balancing is canceled out in the proposed method. Compared with conventional voltage balancing methods, the proposed method possesses good dynamic performance under different working conditions even when the active power is extremely low. Simulation results are provided to validate the effectiveness of the proposed voltage balancing method.

Modular Multilevel Converter used in high- voltage direct current transmission is prevalent nowadays. However, with quantities of IGBT modules, MMC becomes the natural source of Electromagnetic Interference. This paper puts forward a novel strategy called Space Spread-Spectrum-based Carrier Phase-Shifted PWM for MMC to reduce conducted common-mode EMI. This paper analyzes the principle of Space Spread-Spectrum with Double Fourier Transform, which deduces Space Spread-Spectrum-based Carrier Phase-Shifted PWM can change the harmonic from one high amplitude Nf c +nf 0 frequency spike pulse to N low amplitude Nf c +iNΔf+nf 0 frequency harmonics. Space Spread-Spectrum needs to follow several principles. The carrier frequency variation of corresponding submodules of the upper and lower bridge arm should be consistent. Spectrum should spread evenly and the frequency variation should not exceed f c /N. Simulation and experiment shows the good effect of Space Spread-Spectrum, which reduce the conducted common-mode Electromagnetic Interference about 10dB.

Air pollution has witnessed that is an important issue regarding with people health. An excellent air pollution forecasting approach can provide warning to user for avoiding too much polluted air inhalation. Many studies have proposed the model of predicting PM2.5 concentration. How to improve the performance of the forecasting models is an important issue. As well known, ensemble learning is proved that it can improve the forecasting performance. In this paper, in order to improve the performance of air pollution forecasting, we explore an approach named hybrid model framework. which exploit the stacking scheme of ensemble learning. In this work, first we proposed the framework used to forecast the air pollution based on traditional machine approach and deep learning; and exploited Pearson correlation coefficient to calculate the correlation of various models, and finally find out the highest correlation among these models. We conducted experiments and made a comparison with four kinds of models to demonstrate the forecasting performance of hybrid model. The experimental results reveal that the hybrid model is superior than single air pollution forecasting model.

With the advances of industry, air pollution is increasingly becoming serious, and most of governments in the world has deployed many devices to monitor daily air quality. Monitoring and forecasting of air quality has also become an important issue to improve the quality of people's lives. As far as we know, bad air quality does not only affect the health of the respiratory tract, it may but also even cause mental illness. Many researchers have investigated different approaches to work on air quality forecast, and the visualization of forecasting becomes important. In this paper, we present an architecture for visualizing forecasted air quality on a big data platform. We implemented an ETL (Extract-Transform-Load) based framework in the platform, which includes computing nodes and storage nodes. Computational nodes are used for data collection and for air quality forecasting over the next 1 to 8 hours through machine learning and deep learning. Storage nodes are used to retrieve, analyze, and preprocess of collected data. We use the RESTful Web Service as an API, and finally we use the browser to get the data by predefined API and to present the forecasted and monitored results with Google Map API and D3 JavaScript library. It reveals that the visualization on big data framework can work well for air quality analysis.

This paper proposes SocialRank — a content replication protocol that can be used by services deployed in peer-to-peer networks. The proposed solution extends EasyRank replication protocol, by using the knowledge specific to social networks. For this purpose SocialRank includes in the ranking formula a parameter corresponding to the common interests of nodes taking part in the replication. Additionally, expanding the list of nodes that keep replicas was added, and the possibility of withdrawing replicas created during the incomplete replication procedure is provided. The proposed improvements have been a subject to simulation tests, which proved that SocialRank protocol increases the efficiency of the replication. The paper discusses in detail the results obtained during simulation tests, and compares the proposed solution and the EasyRank protocol.

As the advance of technology, it is increasingly speech recognition tools on mobile devices, such as Google Voice and Apple Siri, those have been widely used and have high recognition rate of people's speech. However, these speech recognition tools cannot work well for people with the disease of "Dysphasia" and has very low recognition rate. It is important issue to develop a speech recognition toot for dysphasia, such as Cerebral Palsy(CP) and Amyotrophic Lateral Sclerosis(ALS), to assist those people communicating with others well. Recently, there are various open source programs has been announced, such as Google's Tensorflow, which is used to develop speech recognition based on Deep Neural Networks (DNNs). In this paper, we propose a Convolutional Neural Networks (CNNs) model to perform speech recognition for ALS. The model consists of two hidden CNN layers, each includes one CNN, Rectified Linear Unit (ReLU), Dropout Unit, and MaxPooling Layer. We implement the CNN for dysphasia speech recognition using Google's Tensorflow and collect 33 pronunciations from a dysphasia, each pronunciation at least has 350 training voice files. The highest accuracy is about 63% for one word. And it will be up to totally 94% for top five words. The result shows that it can effectively recognize speech for dysphasia. It can be expected to construct a speech recognition system for assisting dysphasia communicating with others.

With the advance of technology, it is increasingly exhaust emissions have caused air pollution. In particular, PM2.5 (Particulate Matter) has been proven that it has a great correlation with human health. Therefore, the detection and prediction of PM 2.5 air pollution is an important issue. There are countries around the world have built a variety of sensing devices for monitoring PM2.5 concentrations. There were also many studies have been constructed to predict and forecast various air pollution. Therefore, how to accurately forecast PM2.5 becomes an important issue in recent year. In this paper we propose an approach to forecast PM2.5 concentration using RNN (Recurrent Neural Network) with LSTM (Long Short-Term Memory). We exploit Keras, which is a high-level neural networks API written in Python and capable of running on top of Tensorflow, to build a neural network and run RNN with LSTM through Tensorflow. The training data used in the network is retrieved from the EPA (Environmental Protection Administration) of Taiwan from year 2012 to 2016 and is combined into 20-dimensions data; and the forecasting test data is the year 2017. We have conducted experiments to evaluate the forecasting value of PM2.5 concentration for next four hours at 66 stations around the Taiwan. The result shows that the proposed approach can effectively forecast the value of PM2.5.

Air quality monitoring usually rely on statically deploying stationary monitors done by government. However, many air pollution emissions are irregular and uncertain. How to dynamically and effectively monitor air pollution emission will be an important issue for environmental protection. In this paper, we design a LoRa-based air quality monitor tied on Unmanned Aerial Vehicle. It can achieve two goals. The first is that the UAV can real-time send the sensed data back to server if it is flying to long-distance monitored target. The other is that the UAV can perform sensing task itself with minimal human intervention. We implement a prototype of LoRa-based air quality sensor on a UAV and a web-UI for user to configure the route of UAV and view the sensed data immediately. We believe that the prototype can be used easily to monitor the air quality with minimum human intervention.

With the advance of industry, air quality (AQ) is increasingly becoming worse. There are increasingly AQ monitors device have been deployed around country for monitoring air-quality all year long. To estimate and predict AQ, such as PM (particulate matter) 2.5, become an important issue for government to improve people's quality of life. As we can know, there are many factors can affect the AQ, such as traffic, factory exhaust emissions, weather, incineration of garbage, and so on. In most well-developed countries, these pollution sources are monitored for future environmental policy making. In this paper, we will propose a semantic ETL (Extract-Transform-Load) framework on cloud platform for AQ prediction. In the platform, we exploit ontology to concretize the relationship of PM 2.5 from various data sources and to merge those data with the same concept but different naming into the unified database. We implement the ETL framework on the cloud platform, which includes computing nodes and storage nodes. The computing nodes are used to execute data mining algorithms for predicting, and storage modes are used to store retrieved, preprocessed, and analyzed data. We utilize restful web service as the front end API to retrieve analyzed data, and finally we exploit browser to show the visualized result to demonstrate the estimation and prediction. It shows that the big data access framework on the cloud platform can work well for air quality analysis.

For the daily news published on the web, in general, they can be classified into various categories, such as social, politics, entertainment, and so on. These classifications motivate users to watch the desired information. If the classification is wrong, user cannot catch accurately context. How to accurately classify the daily news is becoming an important issue. In this paper, we will propose a method to enhance the effectiveness of news classification. We will utilize the term frequency appeared in variety of classified historical news to training the weighting of each category of each term. And then classify the test news based on the weighting. We propose a framework and an algorithm to training the weighting of each term. The training data, which are over 3500 Chinese news, are collected from UDN and LTN, which are two major electrical news portals in Taiwan. Based on the weighting mechanism, we conduct some experiments to evaluate the effectiveness of the algorithm. The test data are 170 Chinese news, which are collected from Google. The result shows that the traditional manually classification method has up to 13% error classification.

Discount is important in buying behavior and purchasing habits. In this paper, we focus on buying behavior which discount strategy can encourage shopping to devise strategies to boost business owners' sales. This paper introduces a new problem for mining from discounted transaction, and proposes a mining method called the DTM algorithm, which is based on sliding window for maintaining stream transactions. Through the use of this approach, the specific time points at which frequent patterns have a significant increase or decrease in frequency are effectively captured.

Although cloud storage integration services have been developed in recent years that providing a single interface to manage multiple cloud storage accounts across different providers, these cloud storage integration services still remain several issues that do not address. First, because these services do not support arrangement mechanisms for users, they need to keep track of which account they want to upload files to and arrange their storage spaces manually. Second, since each account is divided into different sections, it is impossible to provide a virtual folder that contains files from different cloud storage providers. In addition, when all accounts are nearly full, integration services could not chunk a file to store it across multiple accounts. Third, uploading and moving files across accounts are always done through the integration service's server, which might become a bottleneck when there are a lot of users uploading files at the same time. In this paper, we propose a new cloud storage integration service - Unified Cloud Storage (UCS) for solving the above-mentioned issues, and evaluate its performance by experiments.

In recent years, biofeedback has been widely applied into diagnosis and treatment of various diseases. There are also increasingly research exploiting various ICT (Information & Communication Technology) technologies, such as cloud technology, to achieve diagnosis and treatment. Therefore, how to use mobile cloud technology to assist the disease's diagnosis, to record treatment status, and to infer the result will be an important issue. In this paper, we will propose a mobile cloud platform and framework for the patient of mental illness for evaluating medication response through a variety of biofeedback information collection, integration, and fusion, so that physician can know the patient's situation. The physiological data including Heart Rate Variability and Brain Wave are collected through wearable sensors. And the psychological data is collected through monthly mood chart. The biofeedback physiological and psychological data can be fused into together to show the medication response after patient taking some medicines. An APP for the framework has been developed to show the effectiveness.

ONVIF (Open Network Video Interface Forum) is a popular standard for the video surveillance industries. Since ONVIF specifications do not provide fault-tolerant functions to deal with NVT (Network Video Transmitter) or NVS (Network Video Storage) failure, we propose a fault-tolerant solution for both ONVIF NVT and NVS in this paper. The proposed fault-tolerant extension is also possible to be deployed over cloud-based environments, such as IaaS (Infrastructure as a Service) platforms. The proposed solution is evaluated by the following three performance metrics: the access response time for ONVIF web services, the video recording interrupt duration due to NVS failure, and the video black out time of the surveillance video monitoring system. The experimental results show that all the metrics are within the acceptable range for a real world video surveillance system.

It is shown that the excessive inhalation of PM (Particulate Matter) 2.5 will seriously affect the health of human. Many countries have deployed various detectors for air pollution in order to report concentration of PM2.5 to show how much seriousness of air pollution is. But, what more important is how much PM 2.5 has been inhaled by people anytime and anywhere. Therefore, in this paper, we propose a method for monitoring and estimating people's PM2.5 inhalation by using personally physiological data, such as heart rate (HR), and mobile PM2.5 sensor. We can retrieve people's heart rate from company's cloud platform of smart bracelet and then calculate his/her PM2.5 inhalation based on personal HR and minute ventilation (VE). In the method, we first adopt Bruce protocol to measure various functions of body, such as HR and VE, on a treadmill and exploit regression model to build predictive model for VE. Therefore, we can utilize smart bracelet to monitor people heart rate and then calculate PM2.5 inhalation based on personal HR and mobile PM2.5 sensor (MPS). We think that the method can effectively estimate the PM 2.5 inhalation of physical activity.

In recent years, air quality and environmental protection are getting more and more concern in public, much people are pay attention to air quality of their surroundings. As one of air pollution PM2.5 will be affected to human body directly, it was invisible to the naked eye. According to many epidemiological studies have demonstrated the health hazards of PM2.5, including early death, bronchitis, asthma, cardiovascular disease, lung cancer, and etc. Most of the current PM2.5 Sensor is fixed, and it cannot true reflection of the air quality of people surroundings. Thus, we want to implement Mobile PM2.5 Sensor (MPM) to detect air quality of user surroundings. At the same time, the MPM will send the air quality data to user's smart phone, then send user's GPS location and sensor data to cloud server to construct PM2.5 map by using Google Map API, anyone can get the current air quality information of the region of PM2.5 concentration through PM2.5 maps.

Monthly mood chart is a widely used tool for collecting and evaluating depressed person's mood daily. However, depressed mood is related to high-frequency heart rate variability. Therefore, in order to precisely evaluate personal mood status, it is increasingly important to fuse mood information with heart rate. In this paper1, we design an architecture based on hybrid cloud to collect user's daily mood information that is stored in private cloud and to access heart rate from public cloud; and finally to fuse them together. In addition, we also adopt proof-of-concept prototyping approach to design and implement an APP to view the fused result. It is shown that the architecture is feasible to fuse both together from hybrid cloud.

Due to the emergence of social platforms, people tend to posting their diaries and feeling online for sharing with others. In this paper, we aim to predict whether a user is getting depressed or not through his blog posts on the Internet. For this purpose, we use Latent Dirichlet Allocation (LDA) to find out top frequency words appearing in a user's diaries and use SentiWordNet to calculate the emotion score of the user. Experimental results show that our method is useful in the diagnosis of mental disorder detection in social platforms.