With a large number of possible smart objects in Social Internet Things (SIoT), recommendation system is great necessity to help users find they need. However, traditional techniques usually exploit user's rating or feedback information, which are impractical as such kind user preference information difficult collect the SIoT environment. In addition, temporal context plays an important role object since most tend utilize different at time slots day, e.g., making coffee morning and playing...

To discover the condition of roads, a large number detection algorithms have been proposed, most which apply machine learning methods by time and frequency processing in acceleration velocity data. However, few them pay attention to similarity data itself when vehicle passes over road anomalies. In this article, we propose method detect anomalies comparing windows with various length using Dynamic Time Warping(DTW) method. We model prove that maximum passing through anomaly is linear height...

Multimodal named entity recognition (MNER) for tweets has received increasing attention recently. Most of the multimodal methods used mechanisms to capture text-related visual information. However, unrelated or weakly related text-image pairs account a large proportion in tweets. Visual clues text would incur uncertain even negative effects model learning. In this paper, we propose novel pre-trained based on Relationship Inference and Attention (RIVA) The RIVA controls attention-based with...

While it is well understood that edge computing can significantly facilitate IoT-related applications by deploying servers close to IoT devices, also faces many challenges with numerous devices connected and interacted. One of the most important issues how efficiently deploy under a certain budget explosive growth data scale user base. Existing studies for server placement fail consider user's query preferences since individual users may be interested in events particular regions are keen...

Current scene parsers have effectively distilled abstract relationships among refined instances, while overlooking the discrepancies arising from variations in depth. Hence, their potential to imitate intrinsic 3D perception ability of humans is constrained. In accordance with principle perspective, we advocate first grading depth scenes into several slices, and then digging semantic correlations within a slice or between multiple slices. Two attention-based components, namely Scene Depth...

The energy management information system has became a research hotspot with the rapid development of smart grid, which using for integration micro-grid and traditional electric power grid. However, renewable sources (such as wind energy, tidal etc.) unstable, intermittent controllability characteristics bring number challenges to Solving these problems depend on accurately forecast generation output in certain time. This article outlines tracks main prediction technologies photovoltaic over...

An attention mechanism-based 3D-CNN network was proposed to select the effective bands of hyperspectral images while carrying out model training.

Sarcasm is important to sentiment analysis on social media. Target Identification (STI) deserves further study understand sarcasm in depth. However, text lacking context or missing target makes identification very difficult. In this paper, we introduce multimodality STI and present Multimodal (MSTI) task. We propose a novel multi-scale cross-modality model that can simultaneously perform textual labeling visual detection. the model, extract features enrich spatial information for different...

Proteins are the fundamental biological macromolecules which underline practically all activities. Protein-protein interactions (PPIs), as they known, how proteins interact with other in their environment to perform functions. Understanding PPIs reveals cells behave and operate, such antigen recognition signal transduction immune system. In past decades, many computational methods have been developed predict automatically, requiring less time resources than experimental techniques. this...

Smart cities have drawn a lot of interest in recent years, which employ Internet Things (IoT)-enabled sensors to gather data from various sources and help enhance the quality residents’ life multiple areas, e.g. public safety. Accurate crime prediction is significant for safety promotion. However, complicated spatial-temporal dependencies make task challenging, due two aspects: 1) spatial dependency includes correlations with spatially adjacent regions underlying distant regions, mobility...

In order to effectively predict wind farm power with non-linear and non-stationary characteristics, a prediction model based on empirical mode decomposition (EMD) radial basis function neural networks(RBFNN) was designed. The forecast uses EMD decompose the into several intrinsic functions (IMF) one residue. RBFNN used construct for each IMF component residue, input variables of are triple: speed, direction, history power. All results components were aggregated obtain ultimate result....

Indoor localization is of great importance in pervasive applications and RSS fingerprint known as a quite effective indoor location method. Floor attenuation might not give enough margin discrepancy to classify two neighboring floors, such windows nearby or ring structure. Fingerprint using the nearest Euclidean distance reference point can be interfered by floor. In this paper, multifloor framework with floor identification proposed. The discriminative model trained maximize between-class...

Wi-Fi fingerprinting has become a promising solution for indoor positioning with the rapid deployment of WLAN and growing popularity mobile devices. In fingerprint-based positioning, received signal strengths (RSS) from access points (APs) usually are regarded as fingerprint to label physical location. However, RSS variance caused by heterogeneous devices dynamic environmental status will significantly degrade accuracy. this paper, we first show based on real dataset analyze relation...

Road anomaly detection with crowdsourced sensor data has become an increasingly important field of research over the last few years. Traditional ways for road are either threshold-based techniques or feature-based techniques. However, patterns from crowdsourcing often shifted in time and exhibit local distortions/noise, thus existing methods rely on original greatly limit accuracy detection. In this paper, we present a model by learning scale-invariant features differences between small...

Crime risk prediction is crucial for city safety and residents' life quality. However, without labeled data, it challenging to predict crime in cities. Due municipal regulations maintenance costs, not trivial many cities collect high-quality data. In particular, some have lots of data while others may few. It has been possible develop a model by learning knowledge from with abundant Nevertheless, the inconsistency relevant context between exacerbates difficulty this task. To end, article...

Indoor magnetic-based positioning has attracted tremendous interests in recent years due to its pervasiveness and independence from extra infrastructure. Existing methods for indoor are either point-based fingerprint matching or sequence-based using the raw magnetic field strength. However, magnetometers smartphones vulnerable a few factors such as user's postures walking speed, which causes strength corresponding location often shift time exhibit local distortions, thus greatly limits...

This paper presents a novel passive mobile device localization mode based on IEEE 802.11 Probe Request frames. In this approach, the listener can discover devices by receiving frames and localize them his walking path. The unique location of is estimated geometric diagram right-angled model equations, site-related parameter, that is, path loss exponent, eliminated to make approach site-independent. To implement localization, designed optimal from optional points. performance our method has...

Indoor shopping trajectories provide us with a new approach to understanding user’s behaviour pattern in urban mall, which can be derived from user-generated WiFi logs using indoor localization technology. In this paper, we propose location-aware Point-of-Interest (POI) recommendation service mall that offers user set of POIs by considering both personal interest and location preference. The POI cannot only improve experience but also help the store owner better understand preference intent....