Search

Research on smartphone-based human activity recognition (HAR) is prevalent in the field of healthcare, especially for elderly activity monitoring. Researchers usually propose to use of accelerometers, gyroscopes or magnetometers that are equipped in smartphones as an individual sensing modality for human activity recognition. However, any of these alone is limited in capturing comprehensive movement information for accurate human activity analysis. Thus, we propose a smartphone-based HAR approach by leveraging the inertial signals captured by these three sensors to classify human activities. These heterogeneous sensors deliver information on various aspects of nature, motion and orientation, offering a richer set of features for more accurate representations of the activities. Hence, a deep learning approach that amalgamates long short-term memory (LSTM) in temporal convolutional network (TCN) is proposed. We use independent temporal convolutional networks, coined as temporal convolutional streams, to independently analyse the temporal data of each sensing modality. We name this architecture multi-stream TC-LSTM. The performance of multi-stream TC-LSTM is assessed on the self-collected elderly activity database. Empirical results exhibit that multi-stream TC-LSTM outperforms the existing machine learning and deep learning models, with an F1 score of 98.3 %.

Conventional contact-based hand signature recognition methods are raising hygienic concerns due to shared acquisition devices among the public. Therefore, this research aimed to propose a contactless in-air hand gesture signature (iHGS) recognition method using convolutional recurrent neural networks (C-RNN). Experiments have been conducted to identify the most suitable CNN architecture for the integration of CNN and RNN. A total of four base architectures were adopted and evaluated, namely MS-CNN-A, MS-CNN-B, CNN-A, and CNN-B. Based on the results, CNN-A was selected as the convolutional layer for constructing the integration of C-RNN due to its superior performance, achieving an accuracy rate of 95.15%. Furthermore, three variants of C-RNN were proposed, and experimental results on the iHGS database showed that the ConvBiLSTM achieved the highest accuracy at 98.10%, followed by ConvGRU at 97.47% and ConvLSTM at 97.40%.

Lung cancer is a common cancer in Malaysia, affecting the majority of male citizens. The early detection of lung cancer will decrease its death rate. The only way to detect lung cancer is with a CT scan, and it also requires the doctor to check the scan to confirm the disease. In another way, the computer's support for the detection and diagnosis tool will assist doctors in determining lung cancer more accurately and efficiently. There are three main objectives for this research work. The first target is to study state-of-the-art research work to detect and recognize lung cancer from CT scan images. Then, the article will aim to adopt pre-trained convolutional neural network models in lung cancer detection. It also evaluates the performance of convolutional models on lung cancer imagery data. Then, the pre-trained models with a few added layers and modifications to parameters such as epochs, batch size, optimizer, etc. to conduct model training in this article. After that, Python Pylidc is used in image pre-processing to filter the dataset. Overall, pre-trained models such as ResNet-50, VGG-16, Xception, and MobileNet achieve above-state-of-the-art performance in classifying lung cancer from CT scan images in the range of 78% to 86% accuracy. The best detection accuracy result is the pre-trained VGG-16 model with the addition of some fully connected layers, 16 batch sizes, and the Adam optimizer, which achieved 86.71%.

In the telecom industry, predicting customer churn is crucial for improving customer retention. In literature, the use of single classifiers is predominantly focused. Customer data is complex data due to class imbalance and contain multiple factors that exhibit nonlinear dependencies. In these complex scenarios, single classifiers may be unable to fully utilize the available information to capture the underlying interactions effectively. In contrast, ensemble learning that combines various base classifiers empowers a more thorough data analysis, leading to improved prediction performance. In this paper, a heterogeneous ensemble model is proposed for churn prediction in the telecom industry. The model involves exploratory data analysis, data pre-processing and data resampling to handle class imbalance. In this proposed model, multiple trained base classifiers with different characteristics are integrated through a stacking ensemble technique. Specifically, convolutional-based neural network, logistic regression, decision tree and Support Vector Machine (SVM) are considered as the base classifiers in this work. The proposed stacking ensemble model utilizes the unique strengths of each base classifier and leverages collective knowledge to improve prediction performance with a meta-learner. The efficacy of the proposed model is assessed on a real-world dataset, i.e., Cell2Cell. The empirical results demonstrate the superiority of the proposed model in churn prediction with 62.4% f1-score and 60.62% recall.

Smartphone-based Human Activity Recognition (HAR) identifies human movements using inertial signals gathered from multiple smartphone sensors. Generally, these signals are stacked as one (data-level fusion) and fed into deep learning algorithms for feature extractions. This research studies feature-level fusion, individually processing inertial signals from each sensor, and proposes a lightweight deep temporal learning model, Feature-Level Fusion Multi-Sensor Aggregation Temporal Network (FLF-MSATN), that performs feature extraction on inertial signals from each sensor separately. The raw signals, segmented into equally sized time windows, are passed into individual Dilated-Pooled Convolutional Heads (DPC Heads) for temporal feature analysis. Each DPC Head has a spatiotemporal block containing dilated causal convolutions and average pooling, to extract underlying patterns. The DPC Heads' outputs are concatenated and passed into a Global Average Pooling layer to generate a condensed confidence map before activity classification. FLF-MSATN is assessed using a subject-independent protocol on a publicly available HAR dataset, UCI HAR, and a selfcollected HAR dataset, achieving 96.67% and 82.70% accuracies, respectively. A Data-Level Fusion MSATN is built to compare and verify the model performance attained by the proposed FLF-MSATN. The empirical results show that implementing FLF-MSATN enhances the accuracy by ~3.4% for UCI HAR and ~9.68% for self-collected datasets. [ABSTRACT FROM AUTHOR]

Background: In recent years, human activity recognition (HAR) has been an active research topic due to its widespread application in various fields such as healthcare, sports, patient monitoring, etc. HAR approaches can be categorised as handcrafted feature methods (HCF) and deep learning methods (DL). HCF involves complex data pre-processing and manual feature extraction in which the models may be exposed to high bias and crucial implicit pattern loss. Hence, DL approaches are introduced due to their exceptional recognition performance. Convolutional Neural Network (CNN) extracts spatial features while preserving localisation. However, it hardly captures temporal features. Recurrent Neural Network (RNN) learns temporal features, but it is susceptible to gradient vanishing and suffers from short-term memory problems. Unlike RNN, Long-Short Term Memory network has a relatively longer-term dependency. However, it consumes higher computation and memory because it computes and stores partial results at each level. Methods: This work proposes a novel multiscale temporal convolutional network (MSTCN) based on the Inception model with a temporal convolutional architecture. Unlike HCF methods, MSTCN requires minimal pre-processing and no manual feature engineering. Further, multiple separable convolutions with different-sized kernels are used in MSTCN for multiscale feature extraction. Dilations are applied to each separable convolution to enlarge the receptive fields without increasing the model parameters. Moreover, residual connections are utilised to prevent information loss and gradient vanishing. These features enable MSTCN to possess a longer effective history while maintaining a relatively low in-network computation. Results: The performance of MSTCN is evaluated on UCI and WISDM datasets using subject independent protocol with no overlapping subjects between the training and testing sets. MSTCN achieves F1 scores of 0.9752 on UCI and 0.9470 on WISDM. Conclusion: The proposed MSTCN dominates the other state-of-the-art methods by acquiring high recognition accuracies without requiring any manual feature engineering.

Background Owing to low cost and ubiquity, human activity recognition using smartphones is emerging as a trendy mobile application in diverse appliances such as assisted living, healthcare monitoring, etc. Analysing this one-dimensional time-series signal is rather challenging due to its spatial and temporal variances. Numerous deep neural networks (DNNs) are conducted to unveil deep features of complex real-world data. However, the drawback of DNNs is the un-interpretation of the network's internal logic to achieve the output. Furthermore, a huge training sample size (i.e. millions of samples) is required to ensure great performance. Methods In this work, a simpler yet effective stacked deep network, known as Stacked Discriminant Feature Learning (SDFL), is proposed to analyse inertial motion data for activity recognition. Contrary to DNNs, this deep model extracts rich features without the prerequisite of a gigantic training sample set and tenuous hyper-parameter tuning. SDFL is a stacking deep network with multiple learning modules, appearing in a serialized layout for multi-level feature learning from shallow to deeper features. In each learning module, Rayleigh coefficient optimized learning is accomplished to extort discriminant features. A subject-independent protocol is implemented where the system model (trained by data from a group of users) is used to recognize data from another group of users. Results Empirical results demonstrate that SDFL surpasses state-of-the-art methods, including DNNs like Convolutional Neural Network, Deep Belief Network, etc., with ~97% accuracy from the UCI HAR database with thousands of training samples. Additionally, the model training time of SDFL is merely a few minutes, compared with DNNs, which require hours for model training. Conclusions The supremacy of SDFL is corroborated in analysing motion data for human activity recognition requiring no GPU but only a CPU with a fast- learning rate.

Recognising human activities using smart devices has led to countless inventions in various domains like healthcare, security, sports, etc. Sensor-based human activity recognition (HAR), especially smartphone-based HAR, has become popular among the research community due to lightweight computation and user privacy protection. Deep learning models are the most preferred solutions in developing smartphone-based HAR as they can automatically capture salient and distinctive features from input signals and classify them into respective activity classes. However, in most cases, the architecture of these models needs to be deep and complex for better classification performance. Furthermore, training these models requires extensive computational resources. Hence, this research proposes a hybrid lightweight model that integrates an enhanced Temporal Convolutional Network (TCN) with Gated Recurrent Unit (GRU) layers for salient spatiotemporal feature extraction without tedious manual feature extraction. Essentially, dilations are incorporated into each convolutional kernel in the TCN-GRU model to extend the kernel's field of view without imposing additional model parameters. Moreover, fewer short filters are applied for each convolutional layer to alleviate excess parameters. Despite reducing computational cost, the proposed model utilises dilations, residual connections, and GRU layers for longer-term time dependency modelling by retaining longer implicit features of the input inertial sequences throughout training to provide sufficient information for future prediction. The performance of the TCN-GRU model is verified on two benchmark smartphone-based HAR databases, i.e., UCI HAR and UniMiB SHAR. The model attains promising accuracy in recognising human activities with 97.25% on UCI HAR and 93.51% on UniMiB SHAR. Since the current study exclusively works on the inertial signals captured by smartphones, future studies will explore the generalisation of the proposed TCN-GRU across diverse datasets, including various sensor types, to ensure its adaptability across different applications. [ABSTRACT FROM AUTHOR]

Conventional contact-based hand signature recognition methods are raising hygienic concerns due to shared acquisition devices among the public. Therefore, this research aimed to propose a contactless in-air hand gesture signature (iHGS) recognition method using convolutional recurrent neural networks (C-RNN). Experiments have been conducted to identify the most suitable CNN architecture for the integration of CNN and RNN. A total of four base architectures were adopted and evaluated, namely MS-CNN-A, MS-CNN-B, CNN-A, and CNN-B. Based on the results, CNN-A was selected as the convolutional layer for constructing the integration of C-RNN due to its superior performance, achieving an accuracy rate of 95.15%. Furthermore, three variants of C-RNN were proposed, and experimental results on the iHGS database showed that the ConvBiLSTM achieved the highest accuracy at 98.10%, followed by ConvGRU at 97.47% and ConvLSTM at 97.40%. [ABSTRACT FROM AUTHOR]

This study developed a comprehensive framework using Google Earth Engine to efficiently generate a forest fire inventory dataset, which enhanced data accessibility without specialized knowledge or access to private datasets. The framework is applicable globally, and the datasets generated are freely accessible and shareable. By implementing the framework in Peninsular Malaysia, significant forest fire factors were successfully extracted, including the Keetch–Byram Drought Index (KBDI), soil moisture, temperature, windspeed, land surface temperature (LST), Palmer Drought Severity Index (PDSI), Normalized Vegetation Index (NDVI), landcover, and precipitation, among others. Additionally, this study also adopted large language models, specifically GPT-4 with the Noteable plugin, for preliminary data analysis to assess the dataset's validity. Although the plugin effectively performed basic statistical analyses and visualizations, it demonstrated limitations, such as selectively dropping or choosing only relevant columns for tests and automatically modifying scales. These behaviors underscore the need for users to perform additional checks on the codes generated to ensure that they accurately reflect the intended analyses. The initial findings indicate that factors such as KBDI, LST, climate water deficit, and precipitation significantly impact forest fire occurrences in Peninsular Malaysia. Future research should explore extending the framework's application to various regions and further refine it to accommodate a broader range of factors. Embracing and rigorously validating large language model technologies, alongside developing new tools and plugins, are essential for advancing the field of data analysis. [ABSTRACT FROM AUTHOR]

Customer churn prediction is of utmost importance in the telecommunications industry. Retaining customers through effective churn prevention strategies proves to be more cost-efficient. In this study, attribute selection analysis and deep learning are integrated to develop a customer churn prediction model to improve performance while reducing feature dimensions. The study includes the analysis of customer data attributes, exploratory data analysis, and data preprocessing for data quality enhancement. Next, significant features are selected using two attribute selection techniques, which are chi-square and analysis of variance (ANOVA). The selected features are fed into an artificial neural network (ANN) model for analysis and prediction. To enhance prediction performance and stability, a learning rate scheduler is deployed. Implementing the learning rate scheduler in the model can help prevent overfitting and enhance convergence speed. By dynamically adjusting the learning rate during the training process, the scheduler ensures that the model optimally adapts to the data while avoiding overfitting. The proposed model is evaluated using the Cell2Cell telecom database, and the results demonstrate that the proposed model exhibits a promising performance, showcasing its potential as an effective churn prediction solution in the telecommunications industry. [ABSTRACT FROM AUTHOR]