Depression Intensity Classification from Tweets Using Fast Text Based Weighted Soft Voting Ensemble

Predicting depression intensity from microblogs and social media posts has numerous benefits and applications,including predicting early psychological disorders and stress in individuals or the general public.A major challenge in predicting depression using social media posts is that the existing studies do not focus on predicting the intensity of depression in social media texts but rather only perform the binary classification of depression and moreover noisy data makes it difficult to predict the true depression in the social media text.This study intends to begin by collecting relevant Tweets and generating a corpus of 210000 public tweets using Twitter public application programming interfaces(APIs).A strategy is devised to filter out only depression-related tweets by creating a list of relevant hashtags to reduce noise in the corpus.Furthermore,an algorithm is developed to annotate the data into three depression classes:‘Mild,’‘Moderate,’and‘Severe,’based on International Classification of Diseases-10(ICD-10)depression diagnostic criteria.Different baseline classifiers are applied to the annotated dataset to get a preliminary idea of classification performance on the corpus.Further FastText-based model is applied and fine-tuned with different preprocessing techniques and hyperparameter tuning to produce the tuned model,which significantly increases the depression classification performance to an 84%F1 score and 90%accuracy compared to baselines.Finally,a FastText-based weighted soft voting ensemble(WSVE)is proposed to boost the model’s performance by combining several other classifiers and assigning weights to individual models according to their individual performances.The proposed WSVE outperformed all baselines as well as FastText alone,with an F1 of 89%,5%higher than FastText alone,and an accuracy of 93%,3%higher than FastText alone.The proposed model better captures the contextual features of the relatively small sample class and aids in the detection of early depression intensity prediction from tweets with impactful performances.

Improved organs at risk segmentation based on modified U‐Net with self‐attention and consistency regularisation

Cancer is one of the leading causes of death in the world,with radiotherapy as one of the treatment options.Radiotherapy planning starts with delineating the affected area from healthy organs,called organs at risk(OAR).A new approach to automatic OAR seg-mentation in the chest cavity in Computed Tomography(CT)images is presented.The proposed approach is based on the modified U‐Net architecture with the ResNet‐34 encoder,which is the baseline adopted in this work.The new two‐branch CS‐SA U‐Net architecture is proposed,which consists of two parallel U‐Net models in which self‐attention blocks with cosine similarity as query‐key similarity function(CS‐SA)blocks are inserted between the encoder and decoder,which enabled the use of con-sistency regularisation.The proposed solution demonstrates state‐of‐the‐art performance for the problem of OAR segmentation in CT images on the publicly available SegTHOR benchmark dataset in terms of a Dice coefficient(oesophagus-0.8714,heart-0.9516,trachea-0.9286,aorta-0.9510)and Hausdorff distance(oesophagus-0.2541,heart-0.1514,trachea-0.1722,aorta-0.1114)and significantly outperforms the baseline.The current approach is demonstrated to be viable for improving the quality of OAR segmentation for radiotherapy planning.

Biomechanical Signals Human-Computer Interface for Severe Motor Disabilities

A system that allows computer interaction by disabled people with very low mobility and who cannot use the standard procedure based on keyboard and mouse is presented. The development device uses the patient’s voluntary biomechanical signals, specifically, winks—which constitute an ability that generally remains in this kind of patients—, as interface to control the computer. A prototype based on robust and low-cost elements has been built and its performance has been validated through real trials by 16 users without previous training. The system can be optimized after a learning period in order to be adapted to every user. Also, good results were obtained in a subjective satisfaction survey that was completed by the users after carrying out the test trials.

Introduction to the Special Issue on Recent Advances on Deep Learning for Medical Signal Analysis

Over the past years,deep learning has established itself as a powerful tool across a broad spectrum of domains,e.g.,prediction,classification,detection,segmentation,diagnosis,interpretation,reconstruction,etc.While deep neural networks initially found nurture in the computer vision community,they have quickly spread over medical imaging applications.

PSTCNN: Explainable COVID-19 diagnosis using PSO-guided self-tuning CNN

Since 2019,the coronavirus disease-19(COVID-19)has been spreading rapidly worldwide,posing an unignorable threat to the global economy and human health.It is a disease caused by severe acute respiratory syndrome coronavirus 2,a single-stranded RNA virus of the genus Betacoronavirus.This virus is highly infectious and relies on its angiotensin-converting enzyme 2-receptor to enter cells.With the increase in the number of confirmed COVID-19 diagnoses,the difficulty of diagnosis due to the lack of global healthcare resources becomes increasingly apparent.Deep learning-based computer-aided diagnosis models with high generalisability can effectively alleviate this pressure.Hyperparameter tuning is essential in training such models and significantly impacts their final performance and training speed.However,traditional hyperparameter tuning methods are usually time-consuming and unstable.To solve this issue,we introduce Particle Swarm Optimisation to build a PSO-guided Self-Tuning Convolution Neural Network(PSTCNN),allowing the model to tune hyperparameters automatically.Therefore,the proposed approach can reduce human involvement.Also,the optimisation algorithm can select the combination of hyperparameters in a targeted manner,thus stably achieving a solution closer to the global optimum.Experimentally,the PSTCNN can obtain quite excellent results,with a sensitivity of 93.65%±1.86%,a specificity of 94.32%±2.07%,a precision of 94.30%±2.04%,an accuracy of 93.99%±1.78%,an F1-score of 93.97%±1.78%,Matthews Correlation Coefficient of 87.99%±3.56%,and Fowlkes-Mallows Index of 93.97%±1.78%.Our experiments demonstrate that compared to traditional methods,hyperparameter tuning of the model using an optimisation algorithm is faster and more effective.

Remote sensing and its applications using GNSS reflected signals:advances and prospects

The Global Navigation Satellite Systems(GNSS),including the US’s GPS,China’s BDS,the European Union’s Galileo,and Russia’s GLONASS,offer real-time,all-weather,any-time,anywhere and high precision observations by transmitting L band signals continuously,which have been widely used for positioning,navigation and timing.With the development of GNSS technology,it has been found that GNSS-reflected signals can be used to detect Earth’s surface characteristics together with other signals of opportunity.In this paper,the current status and latest advances are presented on Global Navigation Satellite System-Reflectometry(GNSS-R)in theory,methods,techniques and observations.New developments and progresses in GNSS-R instruments,theoretical modeling,and signal processing,ground and space-/air-borne experiments,parameters retrieval(e.g.wind speed,sea surface height,soil moisture,ice thickness),sea surface altimetry and applications in the atmosphere,oceans,land,vegetation,and cryosphere are given and reviewed in details.Meanwhile,the challenges in the GNSS-R development of each field are also given.Finally,the future applications and prospects of GNSS-R are discussed,including multi-GNSS reflectometry,new GNSS-R receivers,GNSS-R missions,and emerging applications,such as mesoscale ocean eddies,ocean phytoplankton blooms,microplastics detection,target recognition,river flow,desert studies,natural hazards and landslides monitoring.

A Concise Joint Transmit/Receive Antenna Selection Algorithm

For reducing the computational complexity of the problem of joint transmit and receive antenna selection in Multiple-Input-Multiple-Output (MIMO) systems, we present a concise joint transmit/receive antenna selection algorithm. Using a novel partition of the channel matrix, we drive a concise formula. This formula enables us to augment the channel matrix in such a way that the computational complexity of the greedy Joint Transmit/Receive Antenna Selection (JTRAS) algorithm is reduced by a factor of 4n L , where n L is the number of selected antennas. A decoupled version of the proposed algorithm is also proposed to further improve the efficiency of the JTRAS algorithm, with some capacity degradation as a tradeoff. The computational complexity and the performance of the proposed approaches are evaluated mathematically and verified by computer simulations. The results have shown that the proposed joint antenna selection algorithm maintains the capacity perormance of the JTRAS algorithm while its computational complexity is only 1/4n L of that of the JTRAS algorithm. The decoupled version of the proposed algorithm further reduces the computational complexity of the joint antenna selection and has better performance than other decoupling-based algorithms when the selected antenna subset is small as compared to the total number of antennas.

A Compact Flexible Circularly Polarized Implantable Antenna for Biotelemetry Applications

With the help of in-body antennas,the wireless communication among the implantable medical devices(IMDs)and exterior monitoring equipment,the telemetry system has brought us many benefits.Thus,a very thin-profile circularly polarized(CP)in-body antenna,functioning in ISM band at 2.45 GHz,is proposed.A tapered coplanar waveguide(CPW)method is used to excite the antenna.The radiator contains a pentagonal shape with five horizontal slits inside to obtain a circular polarization behavior.A bendable Roger Duroid RT5880 material(εr=2.2,tanδ=0.0009)with a typical 0.25 mm-thickness is used as a substrate.The proposed antenna has a total volume of 21×13×0.25 mm3.The antenna covers up a bandwidth of 2.38 to 2.53 GHz(150 MHz)in vacuum,while in skin tissue it covers 1.56 to 2.72 GHz(1.16 GHz)and in the muscle tissue covers 2.16 to 3.17 GHz(1.01 GHz).GHz).The flexion analysis in the x and y axes was also performed in simulation as the proposed antenna works with a wider bandwidth in the skin and muscle tissue.The simulation and the curved antenna measurements turned out to be in good agreement.The impedance bandwidth of−10 dB and the axis ratio bandwidth of 3 dB(AR)are measured on the skin and imitative gel of the pig at 27.78%and 35.5%,13.5%and 4.9%,respectively,at a frequency of 2.45 GHz.The simulations revealed that the specific absorption rate(SAR)in the skin is 0.634 and 0.914 W/kg in muscle on 1g-tissue.The recommended SAR values are below the limits set by the federal communications commission(FCC).Finally,the proposed low-profile implantable antenna has achieved very compact size,flexibility,lower SAR values,high gain,higher impedance and axis ratio bandwidths in the skin and muscle tissues of the human body.This antenna is smaller in size and a good applicant for application in medical implants.

Coherent ONU based on 850 μm-long cavity-RSOA for next-generation ultra-dense access network

In this Letter, an efficient bidirectional differential phase-shift keying （DPSK）--DPSK transmission for a ultra- dense wavelength division-multiplexed passive optical network is proposed. A single distributed feedback laser at the optical network unit （ONU） is used both as the local laser for downlink coherent detection and the optical carrier for uplink. Phase-shift keying is generated using a low-cost reflective semiconductor optical amplifier （RSOA） at the ONU. The RSOA chip has the bandwidth of 4.7 GHz at the maximum input power and bias current. For uplink transmission, the sensitivity of the RSOA chip reaches -48.2 dBm at the level of bit error rate = 10^-3 for back-to-back, and the penalty for 50 km transmission is less than 1 dB when using polarization diversity.