Search

This research was designed for analyse and compare patterns of EEG signals while blinded and normal people performing touching-leaning. The pattern analysis focuses on important EEG wavebands during touching, such as delta, theta, alpha and gamma waveband. The EEG waveband-datasets were detected and recorded with inexpensive device, the NeuroSky Mindwave, where it is connected to a computer for data analysis through Bluetooth communication and had electric noise reduction chipset inside. The analysis for the EEG waveband pattern-comparisons is performed by utilizing waveband power spectrum and statistical technique based on FFT algorithm, Area Under Curve (AUC), mean, S.D., T-score, and P-value testing. The experiment shown that dominant EEG signal wavebands of blinded people when touching-learning are delta, theta, alpha and gamma. These wavebands were higher than normal people. Moreover, by using statistical analysis T-score and P-values testing, analyzed results illustrate that normal and blinded people EEG wave patterns are significantly different on gamma waveband where blinded people have significantly higher gramma wave during touching-learning activities. The objects used for touching-leaning in the experiments are square, triangle, circle, and hexagon shape of tactile pictures. The observation also shown that blinded people use their muscle in movement more than normal people which also strongly related to gramma wave. In addition, another wave did not relate to statistic significant. This result illustrated to normal and blinded people thinking and imagination with the same pattern.

Currently, the popularity of cosmeceuticals e-commerce is continuously increasing. A competition level among entrepreneurs is higher precisely, especially in the e-commerce platform. As a result, this situation causes to strengthen the competitiveness of several businesses. However, an evaluation of business operation planning which directly associates with the competitiveness of entrepreneurs is still the main issue. Therefore, this study proposes the methodology to define the cosmeceuticals e-commerce competitiveness through SWOT analysis, useful clustering methods including Self-Organizing Map (SOM) and K-means clustering, and Normalized Weight of Criteria for competitiveness criteria evaluation. The SWOT analysis is a reliable algorithm, which can adequately be used to evaluate the business performance with valid questionnaires. It provides the grouped attributes as 4 groups including Strengths, Weaknesses, Opportunities, Threats. Moreover, the appropriated factors result from the previous step will be clustered by SOM and K-means clustering for better data interpretation. SOM calculated 203 instances into 3 clusters; Low, High, No Class with 18 sec for execution time and 94.98% for accuracy. Meanwhile, K-means clustered previous dataset into 2 groups; high performance business and low performance business with 91.33% accuracy. In addition, the clustered data will be specified by Normalized Weight of Criteria for most influential criteria of Strengths, Weaknesses, Opportunities, and Threats factors to accomplish the high Competitiveness level of entrepreneurs. The results from this analysis can definitely help cosmeceuticals entrepreneurs to enhance the business strategy and advanced planning.

Training and development are essential parts of professional development that are necessary for employees to improve their capacity. Generally, professional development program is organized based on personal information, such as background, personal goal, and work experience together with business objectives and job criterion. To promote personalized training in professional development process, the proper classification of individual employee is necessary. This paper thus proposes the classification method for employee classification to promote personalized training in organizations. The machine learning based methods, Decision Tree, Random Forest, and Support Vector Machine, are studied. Synthetic Minority Oversampling Technique (SMOTE) method is used to deal with imbalance data. The open data from Kaggle is used in this paper. For method validation, the data for training and testing are formed into three groups including 80:20, 70:30, and 60:40 respectively. The classification results show that the SMOTE can improve classification performance for all classifiers. Additionally, Random Forest performs the best classification with 96.32% accuracy.

Several countries in the world had a greater number of blinded people. The ways they learn are using sound or tactile pictures. In this paper, authors observed and analyze electroencephalogram signals of them while they learn about the meaning of tactile pictures by inexpensive device. Methods of this paper were collecting dataset by OpenVIBE program, data analyzing by Fast Fourier Transform algorithm and data comparisons by using statistical mean and standard deviation. Results of experiments showed EEG waves, delta, theta, beta and gamma signal during touching were higher than meditation period. Results showed blinded people use more creative, recall and imaginary for learning about pictures. In addition, they used muscle to control touching behaviors.

Digital Technology has been gaining tremendous impact on improving school education quality especially in teaching and learning techniques. In teaching, the process can start from preparing the teaching materials, teaching activities, communicating with students and evaluating student knowledge level. Digital technology can be used to improve and facilitate each stage of this teaching process. However, effectively selecting digital tools for used in each stage can be difficult for teachers especially the ones in non-IT field due to the knowledge of the digital tools, lacking digital skill as well as the limited experiences on how to use them in teaching activities. In many countries, their ministry of education has been providing digital tools and many trainings trying to help teachers to effectively utilize digital tools in their teaching activities; however, the tool selection and the training approach that were used still not effective enough that the tools were used effectively and non-productively in teaching process. This paper reports the result of the developing digital skill for teachers and how they can use digital tools in teaching activities or process. This report is also one part of the major research project supported by the National Broadcasting and Telecommunication Commission (NBTC) of Thailand. The digital tools were collaboratively selected by teachers in the project then trained by researcher team. The tool utilizations in teaching activities are designed, implemented and concluded by community of teachers in the project.

Background and objective Drug-drug interaction (DDI) is one of the main causes of toxicity and treatment inefficacy. This work focuses on non-communicable diseases (NCDs), the non-transmissible and long-lasting diseases since they are the leading cause of death globally. Drugs that are used in NCDs increase the probability of DDIs as a result of long time usage. This work proposes an Integrated Action Crossing (IAC) method that is effective in predicting the NCDs DDIs based on pharmacokinetic (PK) mechanism. Methods Drug-Enzyme (CYP450) and Drug-Transporter actions including substrate, inhibitor and inducer affect the PK mechanism of other drugs. Hence, this paper proposes an enzyme and transporter protein integrated action crossing method for DDIs prediction in NCDs. The NCDs Drugs information was retrieved from the DrugBank database and the actions of enzymes and transporter proteins that were crossed and integrated. The datasets were generated for machine training. Results Three machine learning approaches: Support Vector Machine, k-Nearest Neighbors, and Neural Networks were used for the assessment of the method. Performance evaluation was performed through five-fold cross validation and the different datasets and learning methods were compared. Two layers NNs achieved the best performance at the accuracy of 83.15% (F-Measure 85.23% and AUC 0.901). Conclusions The IAC method delivers better performance compared to the conventional method for the identification of NCDs DDIs.

The Brain Computer Interface (BCI) is the way to interact the human brain to the device or computer. BCI has the several technology for capture the brain wave signals. The electroencephalography (EEG) is the inexpensive way to record the brain signal from the scalp. In this paper, the author used the EEG inexpensive device, NeuroSky MindWave, to connect to the 4 subjects who touching on 2 different types of surface, smooth and rough, and 2 different shapes, sphere and cube. MindWave headset is a BCI device that has 1 electrode which transmit the EEG data to the connected devices. We connected this device to the computer via Bluetooth with MATLAB code and recorded the dataset for 60 seconds and analyzed the wave between 21 to 40 seconds that the time for touching. Finally, we used the mathematic methods, Mean and Standard Deviation for analyze the highest wave values. We discovered the results that the 4 highest EEG bands are Delta, Theta, High-Alpha and Mid-Gamma that affect to the touching behavior. This paper showed the subjects use the creative, higher mental activity and motor function in the touching behavior.

The purposes of this study were to implement the automatic control EC and pH system in reservoir. The automatic control system operate when get signal output from micro-controller. For micro-controller implement EC and pH adjusting equation that were applied from linear regression. The reason to use linear regression because of in experiment amount of A&B solution or nitric acid that adding to reservoir effect to increasing number of EC or pH in the same direction. For EC and pH adjusting equation use 3 variables are target (output), EC or pH measurement (input), and amount of A&B solution or Nitric acid (control process). Micro-controller compute EC and pH adjusting equation to find amount of A&B solution or Nitric acid that need to fill in reservoir for control EC and pH in range of target. The results of experiment show the system that using EC adjusting equation can control EC accuracy more than 80.8% from experiment. And the pH adjusting equation can control pH accuracy more than 95% from experiment.

The expert device is very important within the digital social. This study aims to develop the hardware and algorithm that could work wild area and support any requirement of people's daily life. This device should care your home security-focus on the fire detection. That uses several technologies, to process under the wifi, multi-sensor based on Cloud. Firstly, the sensor calibration to impulse the data detective for quick and ready to process as long time. Next, the FSKF help to filtering data sensor and to estimate the accurate data. The main point of this study is fuzzy logic for fire detection in the home that can send alert messages via an IFTTT process to your smartphone. This is must set warning early-nearly medium range of fire proportion. At last, the OFF-mode help to reduce the power consumption before wifi module send data to the Cloud (ThingSpeak.com). Finally, for the high accuracy of this system to suggest use each sensor type more than one as well as the highly stable.

The expansion and growth of e-commerce business has increased steadily every year. Both original and new operators who are interested in this business because the returns are worth it. But it's not always true that the success of this business. Due to the strategic planning and implementation of business strategy mistakes. Therefore, strategic planning and business strategy, it is necessary and very important. In this paper present the method of selection business strategies planning for the business's e-commerce by using SWOT analysis to help plan and implement strategies in alignment with K-mean method, selection using linear regression for selecting the most appropriate strategy, modeling by using ensemble classifier with Random forest method. Performance measurement this model accuracy is rated over 80 percent through experiments.

Image processing technology is very popular at present. It can be applied to various applications for detecting and processing the digital images. Face detection is a part of image processing. It is used for finding the face of human in a given area. Face detection is used in many applications such as face recognition, people tracking, or photography. In this research, face detection technique is used for detecting and counting the number of passengers in electric vehicle via webcam. The webcam is installed in electric vehicle and connected with Raspberry Pi 2 model B. When electric vehicle leaves from the station, webcam will capture passengers images in the seating area. The images will be adjusted and improved to reduce the noise which is done by software application. The images are sent to the server via 3G communication. Then, the server process the images by using face detection technology and counting the number of passengers in electric vehicle. The system obtains the maximum number of passengers in electric vehicle that process through the images then calculates the seat vacancy of the electric vehicle.

The parking space is one of the crucial problems in the urban city. Most buildings have parking facilities for residents and visitors. The common parking patterns are parallel parking, perpendicular parking and angular parking. Parallel parking is a common design for street side parking whereas perpendicular parking and angular parking are a common design for garages. Angular parking is easy to park than other types. A typical parking space for a personal car is 3 meters in width and 6 meters in length. The parking space have limitation but the number of car increasing along with the economic growth [7]. So the parking space should accommodate the maximum number of car. Normally, the parking space configuration is designed by architects or construction engineers because it's required technical knowledge to calculate a number of car and parking pattern. In parking area, there must be an alleyway for cars to enter and exits. This study is to design and implement a calculation model for number of car and parking pattern in the specified parking area. In the calculation, the dimension and the specific condition of parking area is require. The model calculate the maximum number of car and also calculate the parking pattern for the parking area. This model is use to support non-technical user during the parking area design and can be as engine to develop mobile application for general user in the future.

The next generation of ultra wideband (UWB) wireless technology requirements is antenna performance. Usually the antenna transfer function evaluate are consider used the model of Friis' formula. In case of UWB wireless communication, its can't used Friis' fomula directly for waveform transmission because of UWB impulse radio technology. In this study, evaluate the antennas effect due to UWB transmission waveform by using modified Friis' formula that takes into account the UWB impulse radio transmission, its waveform distortion due to the anatennas with the antenna transfer function and the correlation template at receiver. The wideband antennas are significant pulse-shaping in UWB transmission, the various kinds of the wideband antennas are experimental evaluation, with consideration on this two type wideband antennas effect by using correlator time with received waveform template and the isotropic waveform template.

Most of the wireless communication use wifi technologies such as the IoT which is gaining popularity due to having a smart device. Basically, sensors have been used whereas the various utilization are processed under the analog data and digital data processing. This study aims to design a multi-sensor node. By setting all sensors on the MCU with the GPIO the sensor will only be connected to one processor (MCU). In this connection the invalided multi-sensor data value will appear when the processor startups the following: 1) When the temperature sensor does not work properly 2) The smoke sensor reads an invalid value that is higher than the actual. In this case, we can solve the situation by doing a sensor calibration—setting the calibration time with the dynamic time that follows up on the pinMode() of each sensor and self-calibrate within the FIR filter in part of the setup() portion. 3) When the system is running for a long time it should take the invalid data value, such as high or low from the actual and the value that were swinging. The reason for this case was there were inside and outside noise, or the heat has been collected in the device [14], or noise built in from another device that has been processing for a long time [10][17]. This invalidness could be solved by FSKF to filter and estimate the right data value. Finally, 4) The OFF-Mode should help save power consumption [4] and not sending data sensor all the time—this case is to save your device for working long time and in a long life [12].

Drug-Drug Interactions (DDI) is a cause of treatment inefficacy and toxicity. The most DDI involve drug metabolism which related to enzyme and transporter protein. Drug-enzyme actions that alter the metabolism of other drugs consist of substrate, inhibitor and inducer. Non-communicable diseases (NCDs) are the leading cause of death, drugs that are used in NCDs can increase interaction probability because their long-term usage. This paper proposes Integrated Action Crossing (IAC), a new attribute generation method for DDIs prediction in NCDs. Drugs attributes in NCDs categories were extracted. The actions of enzymes and transporter proteins were crossed for generating dataset for prediction model creation. Neural network (NN) and others machine learning were investigated. Five-fold cross validation was performed for evaluaing the prediction model performance. The results showed that 2 layers NN obtained the best performance of NCDs DDIs prediction model at the accuracy of 83.15%.

A small device with WiFi multi-sensing element is very important under a social digital century. This study aims to implement the hardware and the power of the algorithm with WiFi technologies. Especially, the multi-sensors have to reinforce around a home area and support to any requirement in the term of digital society. This study focus to care the home security— on going to the fire detection with applying several technologies based on a Cloud. Firstly, the multi-sensor calibration has used calibration time and self-calibration as the Finite Impulse Response (FIR). Next, the Full-Scale Kalman Filter (FSKF) helps to fill data and estimate the accuracy data. After that, the fire detection mechanism has used Fuzzy logic to detect and send alert messages over an IFTTT process. There are changed following event-- the data range of fire proportion inside the home. Furthermore, The OFF-Mode has reduced the power consumption suddenly the WiFi module is sent the sensor data to the Cloud. Finally, the WiFi multi-sensor node will use more than one sensor as the same detector will be a high stability and high accuracy.

A WiFi technology was the basis of the Internet of Things (IoT) and many popularity of the wireless communication in the social network. A smart device used any kinds of the detectors all of the analog and digital sensor. This study simulates on the scope of the home security system (HSS). There used three types of sensor; a temperature sensor, smoke-CO, and PIR sensor. This study aims to design a multi-sensor node. All of the sensors are connected on a microcontroller unit (MCU) with the general purpose input output (GPIO). After the connection, there got invalid multi-sensor data. This experiment tried to run over ten times. There appeared some invalid when the processor startup. First, the temperature sensor did not work. Second, the smoke-CO sensors read an invalid value there were higher than the actual. This problem can solve the situation by the sensor calibration methodology—to set the calibration time with the dynamic time follow up on the GPIO function of each sensor and self-calibrate by the finite impulse response (FIR) filter in the part of setup portion. When the system was running for a long time this should take the invalid data. There were high and low from the actual and there got the difference value suddenly a swinging value. During the system was running there had some noise and the heat collected on the device. There got the invalid value. This error is solved by the Full Scale Kalman Filter (FSKF) to fill and estimate the right value. Next, there used the OFF-Mode to save the power consumption and do not send sensor data to the Cloud all time. This method helps the device will be run as long time and work in long life. Finally, there got a high-performance WiFi multi-sensor network.

Drug-Drug Interactions (DDIs) are major causes of morbidity and treatment inefficacy. The prediction of DDIs for avoiding the adverse effects is an important issue. There are many drug-drug interaction pairs, it is impossible to do in vitro or in vivo experiments for all the possible pairs. The limitation of DDIs research is the high costs. Many drug interactions are due to alterations in drug metabolism by enzymes. The most common among these enzymes are cytochrome P450 enzymes (CYP450). Drugs can be substrate, inhibitor or inducer of CYP450 which will affect metabolite of other drugs. This paper proposes enzyme action crossing attribute creation for DDIs prediction. Machine learning techniques, k-Nearest Neighbor (k-NN), Neural Networks (NNs), and Support Vector Machine (SVM) were used to find DDIs for simvastatin based on enzyme action crossing. SVM preformed the best providing the predictions at the accuracy of 70.40 % and of 81.85 % with balance and unbalance class label datasets respectively. Enzyme action crossing method provided the new attribute that can be used to predict drug-drug interactions.

The regression model, approximating the behavior of the utilization rate of WAP's capacity in download-emphasizing network, is estimated based on a 10-client full-factorial experiment with 95.9% adjusted R2. This experiment-based model is further employed to draft a deterministic model estimating the effective per-client throughput, and a probabilistic model approximating the probability to obtain the required throughput. The proposed models can be utilized as easy as the classic model, but the prediction is more precise and the provided probability can be applied to the feasibility analysis, the QOS, and the reliability of systems.