Activity recognition plays an essential role in bridging the gap between low-level sensor data and high-level applications ambient-assisted living systems. With aim to obtain satisfactory rate adapt various application scenarios, a variety of sensors have been exploited, among which, smartphone-embedded inertial are widely applied due its convenience, low cost, intrusiveness. In this paper, we explore power triaxial accelerometer gyroscope built-in smartphone recognizing human physical...

This paper presents a gesture input device, magic wand, with which user can gestures in 3-D space, inertial sensors embedded it generate acceleration and angular velocity signals according to user's hand movement. A trajectory estimation algorithm is employed convert them into on 2-D plane. The recognition based Bayesian networks finds the model maximum likelihood from it. performance of proposed system quite promising; writer-independent rate was 99.2% average for database 15 writers 13 classes.

Community regeneration plays a pivotal role in creating human-centered spaces by transforming spatial configurations, enhancing multifunctional uses, and optimizing designs that promote sustainability vibrancy. However, the influence of such on vitality—particularly its heterogeneity nonlinear effects—remains insufficiently explored. This study presents comprehensive framework combines Difference-in-Differences (DID) method with multiple socio-spatial correlated factors, including place...

The power of end-to-end deep learning techniques to automatically learn latent high-level features from raw signals has been demonstrated in numerous application fields, however, few studies systematically investigate how properly encode the time-series firings binary environment sensors that typically work an event-triggering scheme and have irregular sampling rates for in-home human activity recognition. To this end, we here propose two different methods process streaming sensor readings...

Purpose As a common form of failure in industry, corrosion causes huge economic losses. At present, with the development computational techniques, artificial intelligence (AI) is playing more and important role field scientific research. This paper aims to review application AI protection Design/methodology/approach In this paper, systematically described terms anticorrosion materials methods, image recognition life prediction. Findings With efficient in-depth data processing can rapidly...

We present a 3-D input medium based on inertial sensors for on-line character recognition and an ensemble classification scheme the task. The system allows user to write in air as gesture, with sensor-embedded device held hand. kinds of used are 3-axis accelerometer gyroscope generating acceleration angular velocity signals respectively. For recognition, we technique FDA (Fisher discriminant analysis). tried different combinations sensor test performance. It is also possible estimate 2-D...

Gene selection plays a crucial role in the analysis of microarray data with high dimensionality and small sample size. Incremental wrapper based feature subset (FSS) methods, among various approaches, tend to obtain quality better classification accuracy than filter while it is much more time consuming since interdependence redundancy between features evaluated way. In this paper, we explore introduce Markov Blanket (MB) into incremental FSS process. Rather evaluate all ranked by method, our...

Microarray experiments on gene expression inevitably generate missing values, which impedes further downstream biological analysis. Therefore, it is key to estimate the values accurately. Most of existing imputation methods tend suffer from over-fitting problem. In this study, we propose two regularized local learning for microarray value imputation. Motivated by grouping effect L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>...

Here, a pen-type input device is proposed to track trajectories in 3D space by using accelerometers and gyroscopes. Inertial navigation system (INS) theory, as an autonomous positioning technology widely adopted, employed the users' handwriting motion without external reference sensors or signals. The INS-based system, however, suffers from unbounded error, which grows with time due integration process involved. In order solve this problem, error model built analyze influence of sources....

A new algorithm, the Partial Correlation Statistic (PCS) is presented for structure learning under linear Structural Equation Models. The PCS algorithm can deal with continuous data following arbitrary distribution rather than only a Gaussian distribution. This paper makes two specific contributions. First, arbitrarily distributed datasets, which are generated by structural equation models, if sample size sufficiently large, partial correlation coefficient statistic proved to follow...

Numerical simulation of subcellular [Formula: see text] dynamics with a resolution down to one nanometre can be an important tool for discovering the physiological cause many heart diseases. The requirement enormous computational power, however, has made such simulations prohibitive so far. By using up 12,288 Intel Xeon Phi 31S1P coprocessors on new hybrid cluster Tianhe-2, which is number supercomputer world, we have achieved 1.27 Pflop/s in double precision, brings us much closer...

Early diagnosis significantly improves the survival rate in lung carcinoma patients. This study attempts to construct a predictive network between computational features and semantic of pulmonary nodules using online feature selection causal structure learning. In this paper, we exploit discovery based on streaming algorithm with symmetrical uncertainty algorithm. Different from traditional learning methods that usually obtain all advance then select optimal subset features, proposed...

Existing algorithms of speech-based deception detection are severely restricted by the lack sufficient number labelled data. However, a large amount easily available unlabelled data has not been utilized in reality. To solve this problem, paper proposes semi-supervised additive noise autoencoder model for detection. This updates and optimizes it consists two layers encoder decoder, classifier. Firstly, changes activation function hidden layer network according to characteristics speech....

As the manufacturing industry expands in both scale and energy consumption, challenge of achieving green sustainable development becomes more prominent. One effective approach to this is reducing product consumption by selecting appropriate process parameters. Process parameters fused deposition modeling (FDM) play a crucial role determining during process. Accurately forecasting how these affect essential for realizing FDM. This paper proposes method predicting FDM using Bayesian-optimized...

Feature selection is a significant aspect of speech emotion recognition system. How to select small subset out the thousands data important for accurate classification emotion. In this paper we investigate heuristic algorithm Harmony search (HS) feature selection. We extract 3 sets, including MFCC, Fourier Parameters (FP), and features extracted with The Munich open Speech Music Interpretation by Large Space Extraction (openSMILE) toolkit, from Berlin German database (EMODB) Chinese Elderly...

Feature selection has been playing an important role in analyzing the high-dimension and low-sample-size gene expression profiles towards high classification performance of diseases deep understanding underlying biological mechanisms. Besides performance, stability selected features is another non-ignorable factor evaluating a feature selector, since stable results enhance confidence for true biomarker discovery further validation. In this study, we propose novel method under ensemble...