Search

The use of online food ordering through food systems or applications continues to increase, requiring vendors to implement marketing and sales strategies through surveys, feedback. The problems that arise are building a system analysis model from a collection of tweets with hashtags or usernames for ordering food online . The Support Vector Machine (SVM) algorithm is used for text classification. Tweets are collected into data sets, training data, and testing data, then a classification model of the SVM Algorithm is built. Preprocessing data, tweets are cleansing, tokenized, and stopword remove. From the collected tweets, they are grouped into 10 variables to identify demographic profiles. The results of the analysis are classified as positive sentiments, namely residence, price range, using promos, paid types, halal food while negative sentiments are ethnicity, culture, vegetarianism, place. Classification accuracy is important to validate the results of the SVM model. From 500 train data tweet, the resulting classification is 66% positive sentiment and 34% negative sentiment. Overall accuracy model Linier SVM result 83.2% with accuracy 92.55%.

In March 2020, the Covid-19 outbreak in Indonesia, where the symptoms of the corona virus made the Indonesian people worry and experience depression. It has been almost two years that Covid-19 has not known what causes it, let alone a person's body condition is not good which can result in being attacked by the virus. Covid-19 first appeared in the city of Wuhan, part of China, where it spread very quickly and was deadly. Its spread through direct physical contact with humans is transmitted through the mouth, nose and eyes, therefore a model is needed for the spread of the corona virus. The spread of COVID-19 affects the pattern of interaction between susceptible (susceptible) and infected (infectious) individuals, where human social contact is very heterogeneous and in groups. To influence the impact of the spread of COVID-19 using deep learning approach that is modeled on the spread of COVID-19, individuals are exposed, infected, recover and die. The purpose of this research is to produce good predictions with a deep learning approach for modeling the spread of COVID-19. The results of the deep learning approach for the COVID-19 spread model carried out the 400 time iteration with an MSE achievement of 0.021112.

Stroke is a cardiovascular (CVD) disease caused by the failure of brain cells to get oxygen supply to pose a risk of ischemic damage and result in death. This Disease can detect based on the similarity of symptoms experienced by the sufferer so that early steps can be taking with appropriate counseling and treatment. Stroke detecting requires a machine learning method. In this research, the author used one of the supervised learning classification methods, namely K-Nearest Neighbor (K-NN). K-NN is a classification method based on calculating the distance to training data. This research compares the Euclidean, Minkowski, Manhattan, Chebyshev distance models to obtain optimal results. The distance models have been tested using the stroke dataset sourced from the Kaggle repository. Based on the test results, the Chebyshev model has the highest levels of accuracy compared to the other three distance models with an average accuracy value of 95.49%, the highest accuracy of 96.03%, at K = 10. The Euclidean and Minkowski distance models have the same level of accuracy at each K value with an average accuracy value of 95.45%, the highest accuracy of 95.93% at K = 10. Meanwhile, Manhattan has the lowest average compared to the other distance models, which is 95.42% but has the highest accuracy of 96.03% at the value of K = 6

This chapter will discuss 10 subchapters that will make it more detailed and easier for the reader to master and implement them in their project. Before discussing the subchapters in detail the author discusses the wide use of LCD in various equipment that needs display and the superiority of it compared to the conventional existing displays especially in the low energy consumption of it compare to the rest of the displays, then the author ended this general discussion by mentioning the type of LCD known in the market right now (passive matrix and active matrix). After discussing the LCD in general, the author starts discussing the detailed 10 subchapters. The 10 subchapters are 1. 2 × 16 LCD; 2. LCD controller; 3. LCD instructions; 4. LCD initialization; 5. More instructions; 6. LCD initialization subroutine; 7. Displaying a character on the LCD; 8. Displaying more than 1 character on the LCD; 9. A 4-bit mode 2 × 16 LCD module. To give the readers with a succinct overview of important details or interesting information, the author provides the summary of this chapter in subchapter 10. The author also provided the glossary to enable the readers to quickly study the general terms used in this chapter. Finally, the author provides some questions to enable the reader to test their own knowledge of this chapter. The references is also provided to enable the readers to refer to some articles as the sources of this subchapter and to enable them to enrich their knowledge of this chapter.

Rainfall in Parapat as a part of Lake Toba, North Sumatera and a super-priority tourist destination in Indonesia on May, 13 2021 causes flood, it exceeds Meteorologi, Climatologi and Geophysics Agency’s threshold about intensity of extreme condition. It is related to dynamic of weather’s parameter, especially with convection process and clouds. The landslide material was very closed the road so that the SiantarParapat access was temporarily closed down. Heavy rains that have occurred since noon are said to be the cause of the floods and landslides that occurred at night. This study aims to describe the distribution of rain in Parapat and surrounding areas during floods and landslides. The data used are measured rainfall at the nearest AWS (Automatic Weather Station), and rainfall estimation data from GPM IMERG (Global Precipitation Measurement – The Integrated Multi-SatellitE Retrievals). The results of the study show that convective cloud classification model can be identify the reconstruction of heavy rain that trigger the flood and landslide in parapet. There has been rain with moderate to heavy intensity since 12 May 2021. Heavy rains that have occurred in the last few days have caused soil conditions to become saturated, resulting in landslides. Cloud classification shows the presence of Cumulonimbus cloud clusters that grow and develop around the Lake Toba area before and during heavy rains.

In agriculture, knowing the soil pH is very important not only to determine the right plants in the agriculture land but also to manage plant fertility. The main problem in this paper is how to monitor and control soil potential of hydrogen (pH) using a microcontroller and smartphone. Unfortunately, most of these pH detection systems still use conventional systems. This paper aims to describe the development of an automatic soil pH control system to become an intelligent farmer who can find suitable plants on agricultural land. The novelty of our approach is that sensors and microcontrollers can be used to find out the soil pH of farmland and are routed to smartphones automatically including its location using global positioning system (GPS). To support this research, we use the analog to digital conversion (ADC) signal method. We chose this model because it can be relied upon to detect soil pH precisely. Detection results can be used as a reference to decide also whether to reduce pH or increase soil pH. We have tested not only the type of soil pH in certain areas but also its location is included on Google maps.

Traffic congestion is a problem for almost everyone in big cities. Based on the 2017 traffic condition report released by Inrix, a transportation analysis company, on average, Indonesians were wasting time about 51 hours a year stuck in traffic congestion. Therefore, one of the solutions to overcome this traffic jam problem is by creating an application or system which can find routes with the lowest possible level of a traffic jam from the origin location to the destination. Best First Search algorithm works by selecting the best nodes (with the most economical cost) among other generated nodes from the initial node to the goal node. The route generated by the system will be shown on the map, along with the distance, travel time, algorithm running time, and traffic flow condition of the route. The implementation and testing on the system showed that the distance traveled by walking was less than or equal to the distance traveled by driving. On the other hand, using the same travel mode, the route from origin to destination had different distances and travel time than the vice-versa because of the Best First Search algorithm itself. Nevertheless, in some cases, the distance from the origin to the destination may be the same as from destination to origin because both of them are closed together. The average distance, travel time, and algorithm running time generated from the testing were 2.8 km, 20.375 minutes, and 0.182 seconds. However, the routes generated by the system weren't always optimal because the Best First Search algorithm wasn't taking into account the total travel time taken.

This study comparing the performance of SAW and MOORA methods are in the decision-making process against the evaluation of teacher assistant performance. Then, applying Rank Order Centroid (ROC) is to overcome the process of weighting attributes in decision-making. Subsequently, the data has used by Teaching Assistant Evaluation obtained from UCI Machine Learning Repository, which has 151 instances, five attributes, and 1 class label with multivariate type dataset. The results of testing the calculation of SAW and MOORA methods, obtain the different results where A134 is the best alternative in the calculation of SAW and A132 is the best alternative in the calculation of MOORA. Afterward, if viewing based on running time, the execution process of the SAW method is 0.06 seconds,while the execution process of the MOORA method is 0.02 seconds. In this research, both methods get the same results in ranking, but base on running time, the MOORA method is faster than the SAW method.

In order to optimize the diagnosis of eye disease, the authors want to propose the combination of Support Vector Machine and K-Nearest Neighbor Algorithm to Optimize Diagnosis of Eye Disease. This is a combination method of SVM to get a better classification to diagnose the eye disease. The problem is how to get a better classification to diagnose eye disease. In this research, the hybrid of SVM-KNN is more accurate to classify the eye disease than SVM itself. The accuracy using the combination SVM-KNN algorithm is 94.67%.

This paper proposes and develops an Applying the fuzzy inference method by Mamdani, Tsukamoto and Sugeno. Author will produce a comparison in the final results for each method inthesame case. And this method will be tested on a design model in the process of determining or predicting the export price of a commodity to obtain an optimal profit from the sales of the export commodity. This model includes the prediction of the resale price of exported goods abroad. The process of determining the selling price is calculated using the Fuzzy fuzzyinference method by Mamdani, Tsukamoto, and Sugeno. The variables to be observed are the demand for goods, availability of goods in the farmers, and market price in the foreign countries. The output is an estimated resale price and the profit of each export commodity and the comparison of the selling prices using Mamdani, Tsukamoto, and Sugeno is 2.1% while Mamdani and Sugeno are only 1%after applying logical negation to the basic rules, it is applied to 1.6%. With this comparison, the writer will evaluate the rule base section of Tsukamoto. The results obtained in this study include the prediction model of the resale price of exported goods and the results of the comparison of the performance of the fuzzy interference byMamdani and Tsukamoto logic methods in optimization to provide a more optimal price.

Digital image processing is used to analyze images consisting of visual perceptions whose purpose is to manipulate, modify the image, improve the quality of the image and prioritize the special features contained in the image by using the special attributes of the image characteristics that can be grouped based on the similarity of the image characteristics with each other. One method used in image recognition from artificial neural networks is Learning Vector Quantization (LVQ). The application of the LVQ method for face recognition requires a feature segmentation process so that the feature extraction values obtained only cover the face, this stage is carried out using the Fuzzy C-Means and K-Means Clustering methods. The results of face recognition testing using the LVQ method and K-Means Clustering segmentation obtained an accuracy of 92.22% (training) and 72.5% (testing). Face recognition with Fuzzy C-Means Clustering segmentation obtained face recognition results of 95.78% (training) and 76% (testing). The LVQ-FCM method is better than the LVQ with KMeans, obtained training accuracy of 3.56% and 3.5% for network testing.

Expert Systems were knowledge-based programs that provided expert quality solutions to problems in a specific domain. The concept of this expert system was a computer program that mimics the thought process and expert knowledge in solving a particular problem. The implementation of expert system was widely used in then health field because the expert system was seen as a way of strong expert knowledge in certain field in a computer program. So that decisions could be given in intelligent reasoning. This work used certainty factor (CF) method implemented in a web-based expert system to diagnose various dental and mouth diseases. The experimental result indicated 99% in accuracy of the testing compared compared to the experts diagnosis.

According to the World Health Organization (WHO), it has been recorded that up to now more than 150 million people have diabetes, whether they are elderly people, adults, teenagers, men or women. Early knowledge of diabetes can be seen based on data from patients who already have diabetes. The patient's disease data has previously been stored and arranged in a data warehouse or what is commonly referred to as a dataset. Therefore, it is necessary to process the data contained in the dataset. But the use of data mining techniques themselves must be assisted by using the techniques contained in the data mining, namely classification techniques. K-Nearest Neighbor (K-NN) is one of the methods used in the classification technique. In the results of the classification of the level of confidence obtained in the process, it is seen based on the amount of accuracy. However, there are important issues that need special attention. In the dataset used for the classification process, the data collected contains unbalanced class results (balance). The unbalanced data classification process becomes an important problem, this is because it can cause a decrease in performance. Adaboost is a technique in data mining that can be used to increase the level of accuracy in classification methods. The results showed that the adaboost algorithm can help improve classification performance. This can be seen from the increasing level of accuracy obtained from the process carried out before and after using the adaboost algorithm. The results obtained from the research show that the adaboost algorithm can be used properly to help the performance of the K-Nearest Neighbor algorithm for the classification process on diabetes datasets. It can be seen from 5 tests with values of K = 7, 13, 19, 25 and 31 there is an increase in the accuracy results obtained after using the adaboost algorithm.

Dictionary is a reference that contains words and phrases that are usually sorted alphabetically and is followed by a description of the word or phrase. Dictionary consists of several types, one of which is a dictionary of terms. Psychology dictionary is a dictionary of terms that contain psychological term. Psychology dictionary is identical to the old tome in search and less efficient in its use. Made an Android-based application of psychology dictionary that applies the Horspool algorithm for terms search and SQLite for data storage is a solution. The search algorithm is needed to shorten the time in the search itself. Horspool algorithm apply a saved search information to shift even further because the string matching is done from right to left, so the time to search a string is shorter. Fields used in this research is psychology in general terms. By entering a psychological term as input in the search then we will get the search results. The results of this research show the entire pattern that matches the text.

Penelitian ini dilakukan untuk menganalisis perbandingan hasil kompresi dan dekompresi file audio*.mp3 dan *.wav. Kompresi dilakukan dengan mengurangi jumlah bit yang diperlukan untuk menyimpan atau mengirim file tersebut. Pada penelitian ini penulis menggunakan algoritma Huffman dan Run Length Encoding yang merupakan salah satu teknik kompresi yang bersifat lossless.Algoritma Huffman memiliki tiga tahapan untuk mengkompres data, yaitu pembentukan pohon, encoding dan decodingdan berkerja berdasarkan karakter per karakter. Sedangkan teknik run length ini bekerja berdasarkan sederetan karakter yang berurutan, yaitu hanya memindahkan pengulangan byte yang sama berturut-turut secara terus-menerus. Implementasi algoritma Huffman dan Run Length Encoding ini bertujuan untuk mengkompresi file audio *.mp3 dan *.wav sehingga ukuran file hasil kompresi lebih kecil dibandingkan file asli dimana parameter yang digunakan untuk mengukur kinerja algoritma ini adalah rasio kompresi, kompleksitas yang dihasilkan. Rasio kompresi file audio *.mp3 menggunakan Algoritma Huffman memiliki rata-rata 1.204% sedangkan RLE -94.44%, dan rasio kompresi file audio *.wav memiliki rata-rata 28.954 % sedangkan RLE -45.91%. This research was conducted to analyze the comparison of the results of compression and decompression of *.mp3 and *.wav audio files. Compression was completed by reducing the number of bits needed to save or send the file. In this study, the researcher used the Huffman algorithm and Run Length Encoding which is one of the lossless compression techniques. The Huffman algorithm has three stages to compress data, namely tree formation, encoding and decoding which work based on characters per character. On the other hand, the run length technique works based on a sequence of sequential characters that only move the repetitions of the same byte in succession continuously. The implementation of the Huffman algorithm and Run Length Encoding aimed to compress audio files *.mp3 and *.wav so that the size of the compressed file was smaller than the original file where the parameter used to measure the performance of this algorithm was the compression ratio, and the resulting complexity.*.Mp3 audio file compression ratio using Huffman Algorithm had an average of 1.204% while RLE -94.44%, and compression ratio *.wav audio files had an average of 28.954% while RLE -45.91%.

In this research, we proposed the development of the Document Similarity Detector (DSD) by Jaccard formulation. By this formulation, we can calculate the document similarity based on fitness value. In the retrieval process, we modified the keywords of documents to be the chromosome in the Genetic Algorithm (GA) process in finding the Keyword Solution (KS) by Keyword Competition. Keyword Competition means to eliminate some of keywords based on fitness value of Jaccard. As the data testing, we used the collection paper in Proceeding of SNIKOM, in Medan. We used the GA method to eliminate some of keywords in order to get the most potential keywords of documents in database. Then do the keyword calculation that results in by GA process in a collection of documents in the database. The test results obtained from the similarity of documents selected by users are: 36.11% in Id-paper 165, followed by Id-paper 274 (23.21%), Id-paper 542 (20.67%), Id-paper 172 (14.31%), and Id-paper 189 (12.38%). So it is expected to make it easier for users to find out the similarity of the document.

In hydroponic plants, maintain adequate nutrition are very important to manage fertility of the hydroponic plants. By monitoring nutrient flow, pH stability of nutrients will be easier for a farmer to monitor the hydroponic plants. This paper proposes an Internet of Things (IoT)-based hydroponic plant controlling called ANDES. The novelty of our approach system is that sensor and microcontroller can be built to monitor and control through a smartphone. We show that ANDES can easily configure the connections of various sensors and microcontroller through a smartphone. The developed plant-care intelligence includes automatic pH nutrition, height, temperature and humidity of hydroponics. ANDES intelligence effectively lowers or to higher the pH concentration in the range value of 5.5 - 6.5. In our project, hydroponic nutrient tube was determined that the maximum water level is 6 cm. The sensor will detect if there is a decrease the water height in tube. If the water height has decreased, then the sensor will report and automatically the water pump will turn on in order to increase the nutrition height on the nutrient tube. If the temperature increased more than 28°C in the room of hydroponic plants, then this ANDES tool will turn on the pump automatically to run the cooling water in order to lower room temperature until 22°C.

The development of the amount of data stored through online-based storage systems in cloud computing systems is very large, so there is a possibility that there will be problems in processing such huge data. Therefore in the conference part of the results of this study of course presents an opinion about the new method model that researchers developed is to analyze the data using an integrated analysis model and in measuring the accuracy of large data of course on this occasion using the hierarchy of grid partition (HGP) and as a tool for design that uses the Unified Modeling Language (UML). The big data measurement model in the framework reaches the maximum of the needs of the measuring object entity or which will praise the accuracy of the data. Therefore, there needs to be a further process to support the results of this method model in order to get accurate test results. The working pattern of this big data measurement method model is, of course, agreed on the integration of data stored through the Cloud Server, or Database Management System Server.

In this research, the authors want to propose the combination of Ant Colony Optimization Algorithm and Tabu Search Algorithm as local search to solve Traveling Salesman Problem. This is a hybrid method of ACO to find best routes and get a better running time. One of the classic problems that can be used is TSP. In this research, the authors will compare the hybrid of ACO-TS and ACO. In this research, the hybrid of ACO-TS got the best routes and a better running time than ACO itself. It means that combination of ACO-TS is better than ACO itself. Therefore to get the best routes and a better running time, the author suggested the ACO-TS algorithm to solve TSP.

Skin disease is one of the diseases that are often found in tropical countries like Indonesia. Lack of knowledge about the types and prevention of skin diseases results a person suffering from acute skin diseases. Computer technology is expected to help detect disease early so that it can minimize the occurrence of more dangerous diseases. This paper proposes a method for introducing the shape, color, and texture of skin diseases in digital images and classifying the results of image analysis based on the type of disease in human skin. The method used is a combination of Local Binary Pattern (LBP) and Convolutional Neural Network (CNN) methods which can later be used as sensors or vision for skin diseases automatically. The results of this study can help in the early identification of skin diseases, helping parties who want to know the image value of skin diseases by using LBP and classifying it based on the type of disease using CNN. This study shows the level of accuracy of combining LBP with CNN is quite high with an average value of 92%. In addition, this research can also be used as reference material for the development of further research in image processing that uses LBP and classification using CNN.

This paper discusses fuzzy logic implementation in a load lifting arm robot. Fuzzy logic is used to calculate input values to avoid unsuitable power due to load weight variation. The evaluated performance on the load lifting robot is the successfulness of robot in lifting the load. System should determine the load mass to decide the motor rotation. System is designed by using the ATmega328 microcontroller. Servo motors are employed for driving lever and pinching lever. Mass sensor is employed to predict the load mass. In order to evaluate robot performance, load is limited to a maximum weight of 900 gram due to robot construction. The experiments show better accuracy results in determining the degree of slope servo when Tsukamoto's fuzzy logic model is implemented. Robot is more stable and the average value of the four motor slopes is 34.1%.

The increase of data usage causes problems in data storage, indirectly making the need for data storage also to increase. One alternative solution that can be done is to compress the file so that the file becomes smaller in size so it saves storage space. The algorithm used in this research is the Run Length Encoding algorithm, the Even Rodeh Code algorithm, and the Variable Length Binary Encoding algorithm which are the types of lossless compression. The algorithm will calculate its performance based on Compression Ratio, Ratio of Compression, Redundancy, Compression Time, and Decompression Time. The file that will be used in the data compression process is the file extension *.txt. This study used homogeneous strings (strings that have the same character) and heterogeneous strings (strings that have different characters) in testing the algorithm. In the compression process with a homogeneous string, the combination of the Run Length Encoding algorithm with the Variable Length Binary Encoding algorithm is better than the combination of the Run Length Encoding algorithm and the Even Rodeh Code algorithm with a Compression Ratio of 18.84% and a decompression time of 0.01295 ms. While the compression process on heterogeneous strings from the combination of the Run Length Encoding algorithm with the Even Rodeh Code algorithm is better than the combination of the Run Length Encoding algorithm with Variable Length Binary Encoding algorithms with Compression Ratio of an average of 52.45% and fewer decompression times of 4.93002 ms

Machine learning is an interesting topic nowadays, it is also a subarea of artificial intelligence. An interesting thing that needs to be done in the beginning is how to configure the Development Environment (DE). In this paper, DE is configured using Acer-Notebook with jupyter notebook and Smartphone with google colab for image classification using convolution neural network method. The configuration process is easier to do on Smartphone Realme-C2 than Acer-Notebook and the best level of accuracy is epoch 100 in training and testing data obtained by Smartphone Realme-C2 are 0.70 and 0.66, while Acer-Notebook is 0.70 and 0.63.

Artificial intelligence is a branch of computer science that has advanced quite rapidly and is very useful in helping to alleviate human tasks. One of the models or algorithms part of artificial intelligence is the Convolutional Neural Network (CNN). This model is proven to be very accurate in image data processing. CNN has been used for facial recognition and image data detection systems. In this study, researchers will optimize CNN to obtain a faster and more accurate model for processing plant disease image data in this study using oil palm plant sample data. With the Research and Development method and FGD (Focus Group Discussion) with various experts, this research is expected to produce a CNN model that is faster and more accurate.

This research is motivated by the incomplete use of quantum in existing learning algorithms, so that the proposed learning algorithm is not optimal. Research (Fahri & Neven, 2018) shows that the proposed method of architectural form still uses classical architecture but inputs, weights and targets already use a quantum approach. Based on the results of previous studies, it shows that quantum computing is better than classical computation. Many researchers use quantum computing in the proposed learning algorithm. The model proposed is a quantum circuit architecture with the quantum perceptron method consisting of a quantum bit gate that uses a quantum computational approach as the architecture of the quantum perceptron learning algorithm. Then the authors conduct training and testing of the proposed quantum circuit architecture to test the quantum circuit model that the author proposes. The result of this research is a quantum circuit model with the quantum perceptron method which can be used to solve the learning optimization problem by using a quantum circuit architecture with 5 measurement measurements to show error training and testing = 0, with 9 measurements showing an error training of 1.13%, error testing 2.06%.

Nowadays heart attack is becoming a very deadly disease of human. This attack cannot be predicted, but it can be known from the human heartbeat which is a very vital health parameter of the human cardiovascular system. Heart rate reflects the health conditions of the human cardiovascular system that determined by such as stress at work, before or after sports and the psychology factor. Unfortunately, some people do not know their heart rate before or after doing an activity. This paper proposes of measuring technique of heart rate by using pulse sensor, Arduino microcontroller, and Android Smartphone. It is based on the principle of measuring the variation of blood volume in our body using a light source and detector. We also measure the heart rate by using the ECG or EKG (electrocardiogram) waveform as the comparison result of the pulse sensor with ECG waveform. The sensor consists of an infrared light-emitting-diode (LED) and a photodiode. The LED transmits an infrared light into the fingertip which is reflected back from the blood inside of finger arteries. The results show that this tool can detect the value of the heart rate and is displayed on the screen of an Android. The sensor takes under 10 seconds to detect the value of the heart rate. Notification via SMS will be sent the heart rate such as under normal conditions (BPM, Beat Per Minute 100).

At this time the development of technology that forces everything must move fast, the efficiency of time becomes a measure of the success of a distribution or production. the continuity of production and the success of production itself is related to distribution. When the distribution fails, production will fail. Indonesian people especially in North Sumatra often consume snacks when free time and snacks are the most consumed food by the Indonesian population until now the production of snacks in Indonesia rose to 50% in 2019, causing the food production to have a concept in timely distribution without experiencing delays in distributing goods that have been produced. One of the expeditions in Indonesia experienced problems in terms of late distribution of goods up to 15-20%, this would have a bad impact on the industry in the future, so to overcome the problems above the application of the ant colony optimization algorithm can help the above problems by placing 4 ants and the ants will spread from point A to point H. In the results of this study, there are 4 ants that move together where the best results are obtained by ant 2, closest to the point H, where the value obtained reaches 0,00015

Internet of Things (IoT) collects and sends data over TCP/IP network for the intended application. Some applications designed sensors to be lost after battery exhausted. Therefore, battery lifetime is important. On the other hand, data quality can be disrupted by the harsh environment and interfering noises, so that the transmitted data may be invalid. Transmitting invalid data may exhaust battery sooner, while implementing validation algorithm may absorb more power. This paper proposed a simple peak detection method in IoT sensor to avoid transmitting invalid data. The experimental evaluation was set up and the method was compared to server-based support vector machine classifier. The results show that the proposed method is able to cancel 86% error data transmission with 1.25% power savings.

Study of brain activity generally raises the difficult of distinguishing between the real activity and the artifacts which is caused by an external influence. Therefore, the classification accuracy is still limited by unpredictable signal variations due to background noise. In this paper, we propose a method of artifacts reduction using robust principal component analysis which is applied for an emotion extraction to identify the EEG-SSVEP signals which are induced by the four short movie stimuli (i.e., sad, angry, happy, and calm emotions). The proposed method was tested in real EEG-SSVEP records acquired from eight subjects. The average of 81.5% classification accuracy is achieved for each stimuli. The classification result shows that the proposed method can effectively reduce the artifacts from all subjects.