Experimental Comparison Research Articles

Research and Design of BPM Shortwave Time Signal Modulation Technology Based on Chirp

The shortwave time service system is a vital land-based wireless time service solution, serving as a supplement and backup to the global navigation satellite system. It ensures that time users can access reliable timings, especially in extreme situations. However, the current BPM shortwave time service signal in China faces issues such as insufficient anti-interference reception capabilities and poor timing accuracy. This paper capitalizes on the advantages of Chirp signals and explores a new modulation technology for BPM shortwave time signals that is compatible with the existing modulation system. A Dual Chirp Time-Division Combined Modulation (DCTDCM) scheme is proposed for broadcasting two time signals: Coordinated Universal Time (UTC) and Universal Time 1 (UT1). Furthermore, an in-depth study of the receiving method for this scheme is conducted, with detailed design of its parameters. The designed DCTDCM signals offer a spread spectrum gain of 24 dB and a multipath resolution capability of at least 125 μs, significantly enhancing the anti-interference reception and anti-multipath attenuation capabilities of shortwave time signals. As a result, the availability and timing accuracy of shortwave time signals are substantially improved. Finally, extensive comparative experiments on reception performance validate the effectiveness of this approach.

MAR-YOLOv9: A multi-dataset object detection method for agricultural fields based on YOLOv9

With the development of deep learning technology, object detection has been widely applied in various fields. However, in cross-dataset object detection, conventional deep learning models often face performance degradation issues. This is particularly true in the agricultural field, where there is a multitude of crop types and a complex and variable environment. Existing technologies still face performance bottlenecks when dealing with diverse scenarios. To address these issues, this study proposes a lightweight, cross-dataset enhanced object detection method for the agricultural domain based on YOLOv9, named Multi-Adapt Recognition-YOLOv9 (MAR-YOLOv9). The traditional 32x downsampling Backbone network has been optimized, and a 16x downsampling Backbone network has been innovatively designed. A more streamlined and lightweight Main Neck structure has been introduced, along with innovative methods for feature extraction, up-sampling, and Concat connection. The hybrid connection strategy allows the model to flexibly utilize features from different levels. This solves the issues of increased training time and redundant weights caused by the detection neck and auxiliary branch structures in traditional YOLOv9, enabling MAR-YOLOv9 to maintain high performance while reducing the model’s computational complexity and improving detection speed, making it more suitable for real-time detection tasks. In comparative experiments on four plant datasets, MAR-YOLOv9 improved the mAP@0.5 accuracy by 39.18% compared to seven mainstream object detection algorithms, and by 1.28% compared to the YOLOv9 model. At the same time, the model size was reduced by 9.3%, and the number of model layers was decreased, reducing computational costs and storage requirements. Additionally, MAR-YOLOv9 demonstrated significant advantages in detecting complex agricultural images, providing an efficient, lightweight, and adaptable solution for object detection tasks in the agricultural field. The curated data and code can be accessed at the following link: https://github.com/YangxuWangamI/MAR-YOLOv9.

A Small-Scale Object Detection Algorithm in Intelligent Transportation Scenarios

In response to the problem of poor detection ability of object detection models for small-scale targets in intelligent transportation scenarios, a fusion method is proposed to enhance the features of small-scale targets, starting from feature utilization and fusion methods. The algorithm is based on the YOLOv4 tiny framework and enhances the utilization of shallow and mid-level features on the basis of Feature Pyramid Network (FPN), improving the detection accuracy of small and medium-sized targets. In view of the problem that the background of the intelligent traffic scene image is cluttered, and there is more redundant information, the Convolutional Block Attention Module (CBAM) is used to improve the attention of the model to the traffic target. To address the problem of data imbalance and prior bounding box adaptation in custom traffic data sets that expand traffic images in COCO and VOC, we propose a Copy-Paste method with an improved generation method and a K-means algorithm with improved distance measurement to enhance the model’s detection ability for corresponding categories. Comparative experiments were conducted on a customized 260-thousand traffic data set containing public traffic images, and the results showed that compared to YOLOv4 tiny, the proposed algorithm improved mAP by 4.9% while still ensuring the real-time performance of the model.

GTransCYPs: an improved graph transformer neural network with attention pooling for reliably predicting CYP450 inhibitors

State‑of‑the‑art medical studies proved that predicting CYP450 enzyme inhibitors is beneficial in the early stage of drug discovery. However, accurate machine learning-based (ML) in silico methods for predicting CYP450 inhibitors remains challenging. Here, we introduce GTransCYPs, an improved graph neural network (GNN) with a transformer mechanism for predicting CYP450 inhibitors. This model significantly enhances the discrimination between inhibitors and non-inhibitors for five major CYP450 isozymes: 1A2, 2C9, 2C19, 2D6, and 3A4. GTransCYPs learns information patterns from molecular graphs by aggregating node and edge representations using a transformer. The GTransCYPs model utilizes transformer convolution layers to process features, followed by a global attention-pooling technique to synthesize the graph-level information. This information is then fed through successive linear layers for final output generation. Experimental results demonstrate that the GTransCYPs model achieved high performance, outperforming other state-of-the-art methods in CYP450 prediction.Scientific contributionThe prediction of CYP450 inhibition via computational techniques utilizing biological information has emerged as a cost-effective and highly efficient approach. Here, we presented a deep learning (DL) architecture based on GNN with transformer mechanism and attention pooling (GTransCYPs) to predict CYP450 inhibitors. Four GTransCYPs of different pooling technique were tested on an experimental tasks on the CYP450 prediction problem for the first time. Graph transformer with attention pooling algorithm achieved the best performances. Comparative and ablation experiments provide evidence of the efficacy of our proposed method in predicting CYP450 inhibitors. The source code is publicly available at https://github.com/zonwoo/GTransCYPs.

Order-Preserving Pattern Matching with Partition

Order-preserving pattern matching, which considers the relative orders of strings, can be applied to time-series data analysis. To perform a more meaningful analysis of time-series data, approximate criteria for the order-isomorphism are necessary, considering diverse types of errors. In this paper, we introduce a novel approximation criterion for the order-isomorphism, called the partitioned order-isomorphism. We then propose an efficient O(n+sort(m))-time algorithm for the order-preserving pattern matching problem considering the criterion of partition. A comparative experiment demonstrates that the proposed algorithm is more effective than the exact order-preserving pattern matching algorithm.

Real-Time Policy Optimization for UAV Swarms Based on Evolution Strategies

Multi-agent decision-making faces many challenges such as non-stationarity and sparse rewards, while the complexity and randomness of the real environment further complicate policy development. This paper addresses the high-dimensional policy optimization problems of unmanned aerial vehicle (UAV) swarms. By modeling the problem scenario as a Markov decision process, a real-time policy optimization algorithm based on evolution strategy (ES) pre-training is proposed. This approach combines decision-time planning with background planning to evaluate and integrate different sets of policy parameters in a temporal context. In the experimental phase, the policy network is trained using both ES and REINFORCE algorithms on a constructed simulation platform. Comparative experiments demonstrate the effectiveness of using ES for policy pre-training. Finally, the proposed real-time policy optimization algorithm further improves the performance of the swarm by approximately 10% in simulations, offering a feasible solution for adversarial games between swarms and extending the research scope of evolutionary algorithms.

Predicting Portal Pressure Gradient in Patients with Decompensated Cirrhosis: A Non-invasive Deep Learning Model.

A high portal pressure gradient (PPG) is associated with an increased risk of failure to control esophagogastric variceal hemorrhage and refractory ascites in patients with decompensated cirrhosis. However, direct measurement of PPG is invasive, limiting its routine use in clinical practice. Consequently, there is an urgent need for non-invasive techniques to assess PPG. To develop and validate a deep learning model that predicts PPG values for patients with decompensated cirrhosis and identifies those with high-risk portal hypertension (HRPH), who may benefit from early transjugular intrahepatic portosystemic shunt (TIPS) intervention. Data of 520 decompensated cirrhosis patients who underwent TIPS between June 2014 and December 2022 were retrospectively analyzed. Laboratory and imaging parameters were used to develop an artificial neural network model for predicting PPG, with feature selection via recursive feature elimination for comparison experiments. The best performing model was tested by external validation. After excluding 92 patients, 428 were included in the final analysis. A series of comparison experiments demonstrated that a three-parameter (3P) model, which includes the international normalized ratio, portal vein diameter, and white blood cell count, achieved the highest accuracy of 87.5%. In two distinct external datasets, the model attained accuracy rates of 85.40% and 90.80%, respectively. It also showed notable ability to distinguish HRPH with an AUROC of 0.842 in external validation. The developed 3P model could predict PPG values for decompensated cirrhosis patients and could effectively distinguish HRPH.

Bearing defect detection based on the improved YOLOv5 algorithm.

In the field of bearing defect detection, Aiming at the problem of low efficiency in manual inspection and prone to missed detections in scenarios with small target defects, and overlapping targets, an improved YOLOv5-based object detection method is proposed. Firstly, in terms of feature extraction, the C3 modules in the original backbone of YOLOv5 are replaced with the finer-grained Res2Block modules to improve the model's feature extraction ability. Secondly, in terms of feature fusion, a Bidirectional Feature Pyramid Network (BiFPN) is added to the original neck of YOLOv5 to enhance the fusion ability of shallow graphic features and deep semantic features. Finally, the performance of the improved YOLOv5 algorithm is validated through ablation experiments and comparative experiments with other defect detection algorithms, including the Small_obj algorithm the existing method of adding a small target detection head for identifying small target defects. The experimental results demonstrate that the improved YOLOv5 algorithm exhibits high mAP and accuracy in bearing defect detection, enabling more precise identification the types of small target defects on bearings in complex scenarios with multiple coexisting defects and overlapping detection targets, thereby providing valuable reference for practical bearing defect detection.

A Study of Recommendation Methods Based on Graph Hybrid Neural Networks and Deep Crossing

In the face of complex user behavior patterns and massive data, improving the performance of recommender system models is an urgent challenge. Traditional methods often struggle to effectively handle feature interactions and complex user-item relationships. Combining the advantages of graph neural networks and the Deep Crossing network, this paper proposes a recommendation method based on hybrid neural networks with Deep Crossing (Deep Crossing with Graph Convolution and GRU, DCGCN-GRU). First, by constructing the graph structure of users and items, higher-order feature representations are extracted, and node features are updated using a multilayer graph convolution operation. Then, the higher-order features learned by the graph convolution network are spliced and weighted with the original features to form new feature inputs. Next, a Gated Recurrent Unit (GRU) is introduced to capture the inter-feature temporal dynamic relationships and sequence information. Finally, the Deep Crossing model is utilized to learn the interactions between the fused features at multiple levels and enhance the interactions between the features. Comparative experiments on three public datasets, MovieLens-ml-25m, Book-Crossings, and Amazon Reviews’23, show that the model achieves significant improvements in accuracy, mean square error (MSE), and mean absolute error (MAE).

Study on anti-reflection of deep soft and high gas coal seam floor by deep hole controlled blasting

Addressing the practical challenges of difficult drilling for blasting-induced permeability enhancement in deep, soft, and high-gas coal seams, where fractures remain underdeveloped and prone to re-compaction, this study proposes blasting operations within the floor strata. This approach aims to enhance the permeability of soft coal seams, thereby extending the duration of effective gas extraction. A bidirectional loading gas–solid coupling blasting simulation system was established in the laboratory, enabling multi-faceted analysis of experimental models through macroscopic crack patterns, internal damage mechanisms, and strain data of coal and rock masses. Comparative experiments were conducted, contrasting various control hole spacings with conventional blasting techniques. The findings reveal that as the blasting stress wave traverses the control hole walls, tensile stress waves are reflected, facilitating crack propagation. The guiding effect of the control holes and the spatial compensation they provide significantly increase the extension distance of explosion-induced cracks, resulting in directional failure of the test specimens and heightened damage in the far field of the blast. After the blasting process, the arrangement of control holes can result in an increase of up to 133% in damage to the coal seam and a reduction of up to 167% in damage to the floorboard compared to the model without control holes. Notably, when the control holes are proximal to the coal-rock interface, the near-end coal body experiences the most pronounced effects, with peak damage and tensile strain in the d = 20 mm model being 1.93 and 1.79 times higher, respectively, than those in models without control holes. Conversely, for control holes located further from the interface, the distal coal body experiences the greatest influence, exhibiting 1.53 and 1.55 times higher peak damage and tensile strain, respectively, in the d = 80 mm model compared to uncontrolled counterparts. Field observations at the C13 coal seam of a mine within the Huainan mining area corroborate these findings, where the volume of gas extraction and its concentration experienced a rapid increase following blasting and penetration enhancement. Optimum permeability enhancement occurs when the blasting hole is situated 4 m from the extraction point, resulting in a 131% increase in gas extraction purity from 0.15 × 10–3 m3/min to 1.97 × 10–3 m3/min. Furthermore, gas concentration soars by 373%, from 5.86% to 21.86%. These research outcomes offer valuable insights and hold considerable reference significance for blasting-induced permeability enhancement in deep, soft, and high gas coal seams.

Optimizing urban bus network based on spatial matching patterns for sustainable transportation: A case study in Harbin, China.

The rapid economic development and accelerating urbanization have led to a significant mismatch between the urban bus network allocation and the population flow. Therefore, this paper investigates this challenge by exploring the intricate relationship between the population flow dynamics, traffic congestion conditions, and the efficient allocation of bus resources. In response, two key indexes were introduced based on spatial matching patterns to assess the urban bus network: the Population-Bus Match Index evaluates the matching degree between supply and demand, and the Population-Congestion Match Index evaluates the matching degree between utilization and saturation. Additionally, two distinct optimization strategies have been proposed to enhance the urban bus network. The first optimization strategy considers the bus network's current status, while the second aspires to an idealized scenario. Subsequently, the potential contributions of each bus station in reducing CO2 emission reduction after implementing the two optimization strategies are quantified. Utilizing a case study focused on Harbin, the proposed methods are validated. The findings unveil a substantial misalignment between supply and demand within the bus network during peak periods, with nearly half of the bus stations experiencing a disparity between utilization and saturation. Comparative experiments across different optimization strategies reveal that the second optimization strategy significantly outperforms the first, but the first optimization strategy has a higher degree of CO2 emission reduction contribution. The results of this study provide decision-makers with an environmentally oriented vantage point for the discerning selection of optimization strategies and leave valuable insights for urban areas confronting transportation challenges.

Salting-out Effects in Extraction of Phenol from Aqueous Solutions Using a Hydrophobic Deep Eutectic Solvent Based on Dodecanoic and Octanoic Acids: Experimental Study and Comparison

Salting-out Effects in Extraction of Phenol from Aqueous Solutions Using a Hydrophobic Deep Eutectic Solvent Based on Dodecanoic and Octanoic Acids: Experimental Study and Comparison

Improving ncRNA family prediction using Multi-Modal contrastive learning of sequence and structure.

Recent advancements in high-throughput sequencing technology have significantly increased the focus on non-coding RNA (ncRNA) research within the life sciences. Despite this, the functions of many ncRNAs remain poorly understood. Research suggests that ncRNAs within the same family typically share similar functions, underlining the importance of understanding their roles. There are two primary methods for predicting ncRNA families: biological and computational. Traditional biological methods are not suitable for large-scale data prediction due to the significant human and resource requirements. Concurrently, most existing computational methods either rely solely on ncRNA sequence data or are exclusively based on the secondary structure of ncRNA molecules. These methods fail to fully utilize the rich multimodal information available from ncRNAs, thereby preventing them from learning more comprehensive and in-depth feature representations. To tackle these problems, we proposed MM-ncRNAFP, a multi-modal contrastive learning framework for ncRNA family prediction. We first used a pre-trained language model to encode the primary sequences of a large mammalian ncRNA dataset. Then, we adopted a contrastive learning framework with an attention mechanism to fuse the secondary structure information obtained by graph neural networks. The MM-ncRNAFP method can effectively fuse multi-modal information. Experimental comparisons with several competitive baselines demonstrated that MM-ncRNAFP can achieve more comprehensive representations of ncRNA features by integrating both sequence and structural information. This integration significantly enhances the performance of ncRNA family prediction. Ablation experiments and qualitative analyses were performed to verify the effectiveness of each component in our model. Moreover, since our model is pre-trained on a large amount of ncRNA data, it has the potential to bring significant improvements to other ncRNA-related tasks. MM-ncRNAFP and the datasets are available at https://github.com/xuruiting2/MM-ncRNAFP.

WiTUnet: A U-shaped architecture integrating CNN and Transformer for improved feature alignment and local information fusion

Low-dose computed tomography (LDCT) has emerged as the preferred technology for diagnostic medical imaging due to the potential health risks associated with X-ray radiation and conventional computed tomography (CT) techniques. While LDCT utilizes a lower radiation dose compared to standard CT, it results in increased image noise, which can impair the accuracy of diagnoses. To mitigate this issue, advanced deep learning-based LDCT denoising algorithms have been developed. These primarily utilize Convolutional Neural Networks (CNNs) or Transformer Networks and often employ the Unet architecture, which enhances image detail by integrating feature maps from the encoder and decoder via skip connections. However, existing methods focus excessively on the optimization of the encoder and decoder structures while overlooking potential enhancements to the Unet architecture itself. This oversight can be problematic due to significant differences in feature map characteristics between the encoder and decoder, where simple fusion strategies may hinder effective image reconstruction. In this paper, we introduce WiTUnet, a novel LDCT image denoising method that utilizes nested, dense skip pathway in place of traditional skip connections to improve feature integration. Additionally, to address the high computational demands of conventional Transformers on large images, WiTUnet incorporates a windowed Transformer structure that processes images in smaller, non-overlapping segments, significantly reducing computational load. Moreover, our approach includes a Local Image Perception Enhancement (LiPe) module within both the encoder and decoder to replace the standard multi-layer perceptron (MLP) in Transformers, thereby improving the capture and representation of local image features. Through extensive experimental comparisons, WiTUnet has demonstrated superior performance over existing methods in critical metrics such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity (SSIM), and Root Mean Square Error (RMSE), significantly enhancing noise removal and image quality. The code is available on github https://github.com/woldier/WiTUNet.

Research on efficient matching method of coal gangue recognition image and sorting image

When the coal gangue sorting robot sorts coal gangue, the position of the target coal gangue will change due to belt slippage, deviation, and speed fluctuations of the belt conveyor. This will cause the robotic to fail in grasping or miss grasping. We have developed a solution to this problem: the IMSSP-Net two-stage network gangue image fast matching method. This method will reacquire the target gangue position information and improve the robot’s grasping precision and efficiency. In the first stage, we use SuperPoint to guarantee the scene adaptability and credibility of feature point extraction. We have enhanced Superpoint’s ability to detect feature points further by using the improved Multi-scale Retinex with Color Restoration enhancement algorithm. In the second stage, we introduce SuperGlue for feature matching to improve the robustness of the matching network. We eliminated erroneous feature matching point pairs and improved the accuracy of image matching by adopting the PROSAC algorithm. We conducted image matching comparison experiments under different object distances, scales, rotation angles, and complex conditions. The experimental platform adopts the double-manipulator truss-type coal gangue sorting robot independently developed by the team. The matching precision, recall, and matching time of the method are 98.2%, 98.3%, and 84.6ms, respectively. The method can meet the requirements of efficient and accurate matching between coal gangue recognition images and sorting images.

Organic cropping systems balance environmental impacts and agricultural production

Agriculture provides food to a still growing population but is a major driver of the acceleration of global nutrient flows, climate change, and biodiversity loss. Policies such as the European Farm2Fork strategy aim to mitigate the environmental impact of land-use by fostering organic farming. To assess long-term environmental impact of organic food production we synthesized more than four decades of research on agronomic and environmental performance of the oldest system comparison experiment on organic and conventional cropping systems (DOK experiment). Two organic systems (bioorganic and biodynamic) are compared with two conventional (manure-based integrated and mineral-based) systems all with the same arable crop rotation including grass-clover, and manure from livestock integrated in all except the mineral-based system. Organic systems used 92% less pesticides and 76% less mineral nitrogen than conventional systems. Nitrogen use efficiency, that also considers biological nitrogen fixation, was above 85% for all systems. Organic fertilization with farmyard manure maintained or increased soil carbon and nitrogen stocks in the long term, especially in the biodynamic system with manure compost. Conventional mineral-based cropping reduced soil organic carbon and nitrogen stocks. Higher soil organic carbon stocks in organic cropping did not translate to increased N2O emissions, which were the main driver for 56% lower soil-based, area-scaled climate impact compared to the integrated conventional system with manure. Organic cropping systems, especially compost-based biodynamic, showed enhanced soil health, richness of micro- and macrofauna and weed species. Highest yields were achieved in integrated conventional system, with highest total nitrogen inputs and enhanced soil health compared to pure mineral fertilization. Yet, these benefits come at the cost of lower nitrogen use efficiency and higher N2O emissions. Despite a rigorous reduction of inputs yields of the organic systems achieved 85% of the conventional systems. We demonstrate at field level that organic cropping systems with reduced external nutrient inputs have less climate impact and a larger in-situ biodiversity, while providing a fertile ground for the future development of sustainable agricultural production systems.

Personalized movie recommendation in IoT-enhanced systems using graph convolutional network and multi-layer perceptron

Exploring the optimization of communication strategies for animation films in the context of cross-cultural communication, this research integrates the Internet of Things (IoT) and convolutional networks. The research constructs a collaborative filtering (CF) movie recommendation model based on a graph convolutional neural network (GCN) and investigates its application in cross-cultural communication. The fusion of IoT and convolutional networks in movie communication is also analyzed, and the effectiveness of the proposed GCN-CF model is validated through comparative experiments. The results indicate that, compared to other models, the GCN-CF model achieves the lowest Root Mean Square Error (RMSE) on the MovieLens 100 K and MovieLens 1 M datasets, with values of 0.8762 and 0.8275, respectively. Compared to traditional models, the GCN-CF model exhibits significantly superior performance in terms of RMSE, with reductions ranging from 0.6 to 5.2%, highlighting its heightened detection accuracy and overall performance. Moreover, the performance of the GCN-CF model is enhanced after introducing attention mechanisms and auxiliary information on both datasets, showing an improvement of 0.4% compared to the scenario without these additions. This data demonstrates the effectiveness of attention mechanisms and auxiliary information. Finally, the research presents an animation film communication strategy based on IoT and convolutional networks, offering novel insights for film production and communication, along with positive implications for cultural exchange and the advancement of the global media industry.

A rolling bearing fault signal denoising algorithm that combines a new adaptive information entropy with a new wavelet threshold function

Abstract Mechanical fault diagnosis is of great significance to industrial automation, and extracting vibration fault signals is one of the important tasks in mechanical health monitoring and fault diagnosis. However, due to the complex working environment of rolling bearings, a large amount of noise makes it difficult to extract vibration fault signals. Denoising the vibration signal of bearings can remove interference noise, simplify the early identification of signal features, and thus improve diagnostic accuracy and mechanical maintenance efficiency. This paper proposes a rolling bearing fault signal denoising algorithm, which constructs a new feature extraction and denoising function. This method first decomposes the noisy signal into Intrinsic Mode Functions (IMFs) by Intrinsic Computing Expressive Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN). Secondly, a new adaptive information entropy threshold function is constructed to extract the noisy IMFs from it. Then, the noisy IMF signal is denoised by a new wavelet threshold function. Finally, the noise-free IMFs are reconstructed to reconstruct the denoised signal. To verify the actual performance of the algorithm, comparative experiments were conducted on a self-collected dataset and a public dataset, the results show that this method improves the continuity of signal reconstruction and can remove various types of vibration fault noise signals more effectively and accurately, \hl{thereby improving the fault detection accuracy by 2%-9%.

A Normal Displacement Model and Compensation Method of Polishing Tool for Precision CNC Polishing of Aspheric Surface

The position accuracy of the polishing tool affects the surface quality of the polished aspheric surface. The contact deformation among the polishing tool, abrasives, and aspheric part can cause a displacement, which, in turn, will cause a position error of the polishing tool, which will lead to a significant change in the polishing force. In order to resolve this error, this paper proposed a method of normal displacement compensation for a computer numerical controlled (CNC) polishing system by controlling the polishing force. Firstly, the coupling principle between the polishing force and the position of the polishing tool is expounded, and the relationship between normal displacement and deformation is analyzed. Based on Hertz’s theory, a model of normal displacement is established. Then, on the basis of the decoupled polishing system developed, a normal displacement compensation method was proposed. Finally, a group of comparative experiments was carried out to verify the effectiveness of the proposed method. Compared with no displacement compensation, when the part was polished with the normal displacement compensation method, the value of roughness decreased from 0.4 µm to 0.21 µm, and the unevenness coefficient of surface roughness decreased from 112.5% to 19%. The experimental results show that the polishing quality is improved greatly, and the aspheric surfaces can be polished more uniformly with the method proposed in this paper.

Unsupervised selection of acoustic indices: An experimental comparison for characterizing unlabeled audio recordings from sub-Andean forest soundscapes

We present an experimental comparison of applying six unsupervised (i.e. not relying on class labels) and almost parameterless feature selection methods for ranking acoustic indices. The study is aimed at guiding the practitioner in choosing appropriate acoustic indices when, in absence of class labels, a small but meaningful number of features to characterize soundscapes is desired. Forty acoustic indices were considered, which correspond to seventeen temporal, spectral and soundscape features, along with their basic statistics. Three publicly available soundscape datasets, registered in a sub-Andean forest, were used for the experiments; moreover, several subsets were considered according to the different times of the day. Results reveal that the Acoustic Evenness Index is the most important one in terms of representational power according to the six considered selection criteria, followed by the Acoustic Complexity Index when conditions are relaxed to examine features ranked among the top-five. Besides, the Bioacoustic Index, the Acoustic Diversity Index and the Root Mean Square Energy stand out as important features when characterizing days in their separated parts.