Deep Forest Algorithm Research Articles

Golgi_DF: Golgi proteins classification with deep forest.

Golgi is one of the components of the inner membrane system in eukaryotic cells. Its main function is to send the proteins involved in the synthesis of endoplasmic reticulum to specific parts of cells or secrete them outside cells. It can be seen that Golgi is an important organelle for eukaryotic cells to synthesize proteins. Golgi disorders can cause various neurodegenerative and genetic diseases, and the accurate classification of Golgi proteins is helpful to develop corresponding therapeutic drugs. This paper proposed a novel Golgi proteins classification method, which is Golgi_DF with the deep forest algorithm. Firstly, the classified proteins method can be converted the vector features containing various information. Secondly, the synthetic minority oversampling technique (SMOTE) is utilized to deal with the classified samples. Next, the Light GBM method is utilized to feature reduction. Meanwhile, the features can be utilized in the penultimate dense layer. Therefore, the reconstructed features can be classified with the deep forest algorithm. In Golgi_DF, this method can be utilized to select the important features and identify Golgi proteins. Experiments show that the well-performance than the other art-of-the state methods. Golgi_DF as a standalone tools, all its source codes publicly available at https://github.com/baowz12345/golgiDF. Golgi_DF employed reconstructed feature to classify the Golgi proteins. Such method may achieve more available features among the UniRep features.

Modelling the furnace temperature field of a municipal solid waste incinerator using the numerical simulation and the deep forest regression algorithm

The furnace temperature field is key information to reflect the combustion conditions and is related to process safety, equipment failure rates, and pollution emissions in municipal solid waste incinerators. However, the measurement of the actual temperature distribution of the incinerator is difficult because of the limitations of physical sensors (thermocouples). A new furnace temperature field modeling strategy was proposed to address the above problem. It included five virtual temperature points, with the numerical simulation and the deep forest regression algorithm used. First, numerical models of different benchmark conditions were established via FLIC. Second, modeling data were obtained by orthogonal experiments. Third, deep forest regression based on the fully connected cross layer (DFR-clfc) was used to establish five virtual temperature models. The five models had high prediction performance (R2 = 0.707, 0.870, 0.918, 0.895, and 0.907) and characterized the distribution of the furnace temperature. The grate speed had the greatest influence on temperature distribution, and based line condition (2) was more suitable for a stable operation. The novelties of the work are as follows. (1) Furnace temperature field estimation based on FLIC simulation and data-driven model. (2) Multi-conditions design based on micro, composition, and operating parameters. (3) Multi-dimensional visual analysis of incinerator virtual temperature field. (4) Temperature field model by using deep ensemble forest with potential interpretability.

Accelerating the Discovery of Anticancer Peptides through Deep Forest Architecture with Deep Graphical Representation

Cancer is one of the leading diseases threatening human life and health worldwide. Peptide-based therapies have attracted much attention in recent years. Therefore, the precise prediction of anticancer peptides (ACPs) is crucial for discovering and designing novel cancer treatments. In this study, we proposed a novel machine learning framework (GRDF) that incorporates deep graphical representation and deep forest architecture for identifying ACPs. Specifically, GRDF extracts graphical features based on the physicochemical properties of peptides and integrates their evolutionary information along with binary profiles for constructing models. Moreover, we employ the deep forest algorithm, which adopts a layer-by-layer cascade architecture similar to deep neural networks, enabling excellent performance on small datasets but without complicated tuning of hyperparameters. The experiment shows GRDF exhibits state-of-the-art performance on two elaborate datasets (Set 1 and Set 2), achieving 77.12% accuracy and 77.54% F1-score on Set 1, as well as 94.10% accuracy and 94.15% F1-score on Set 2, exceeding existing ACP prediction methods. Our models exhibit greater robustness than the baseline algorithms commonly used for other sequence analysis tasks. In addition, GRDF is well-interpretable, enabling researchers to better understand the features of peptide sequences. The promising results demonstrate that GRDF is remarkably effective in identifying ACPs. Therefore, the framework presented in this study could assist researchers in facilitating the discovery of anticancer peptides and contribute to developing novel cancer treatments.

Using deep ensemble forest for high-resolution mapping of PM2.5 from MODIS MAIAC AOD in Tehran, Iran

High-resolution mapping of PM2.5 concentration over Tehran city is challenging because of the complicated behavior of numerous sources of pollution and the insufficient number of ground air quality monitoring stations. Alternatively, high-resolution satellite Aerosol Optical Depth (AOD) data can be employed for high-resolution mapping of PM2.5. For this purpose, different data-driven methods have been used in the literature. Recently, deep learning methods have demonstrated their ability to estimate PM2.5 from AOD data. However, these methods have several weaknesses in solving the problem of estimating PM2.5 from satellite AOD data. In this paper, the potential of the deep ensemble forest method for estimating the PM2.5 concentration from AOD data was evaluated. The results showed that the deep ensemble forest method with [Formula: see text] gives a higher accuracy of PM2.5 estimation than deep learning methods ([Formula: see text]) as well as classic data-driven methods such as random forest ([Formula: see text]). Additionally, the estimated values of PM2.5 using the deep ensemble forest algorithm were used along with ground data to generate a high-resolution map of PM2.5. Evaluation of produced PM2.5 map revealed the good performance of the deep ensemble forest for modeling the variation of PM2.5 in the city of Tehran.

A novel method for predicting DNA N4-methylcytosine sites based on deep forest algorithm.

N4-methyladenosine (4mC) methylation is an essential epigenetic modification of deoxyribonucleic acid (DNA) that plays a key role in many biological processes such as gene expression, gene replication and transcriptional regulation. Genome-wide identification and analysis of the 4mC sites can better reveal the epigenetic mechanisms that regulate various biological processes. Although some high-throughput genomic experimental methods can effectively facilitate the identification in a genome-wide scale, they are still too expensive and laborious for routine use. Computational methods can compensate for these disadvantages, but they still leave much room for performance improvement. In this study, we develop a non-NN-style deep learning-based approach for accurately predicting 4mC sites from genomic DNA sequence. We generate various informative features represented sequence fragments around 4mC sites, and subsequently implement them into a deep forest (DF) model. After training the deep model using 10-fold cross-validation, the overall accuracies of 85.0%, 90.0%, and 87.8% were achieved for three representative model organisms, A. thaliana, C. elegans, and D. melanogaster, respectively. In addition, extensive experiment results show that our proposed approach outperforms other existing state-of-the-art predictors in the 4mC identification. Our approach stands for the first DF-based algorithm for the prediction of 4mC sites, providing a novel idea in this field.

A data-driven fault detection approach for Modular Reconfigurable Flying Array based on the Improved Deep Forest

The Modular Reconfigurable Flying Array, as a kind of special modular rotorcraft UAV, can change its topology configurations fitting for different tasks and varying work scenarios. However, strongly nonlinear characteristics and limited uncertainties of Modular Reconfigurable Flying Array make its sensor faults happen easily and fail to be detected with higher accuracy. In the light of the potential serious losses of its sensors faults, it is valuable and challenging to detect the sensor faults of the aircraft accurately and effectively. Therefore, a data-driven multivariate regression approach based on the Improved Deep Forest is proposed to fulfill sensor faults detection. Firstly, the Deep Forest algorithm is improved by adding the enhanced cascade layer structure and redesigning the inter-layer loss function to pursuit better prediction accuracy. And then, the Improved Deep Forest algorithm is used to establish a multivariate regression model and obtains an estimation of the monitored parameter by learning the historical flight data. Finally, the residual between the actual flight data and the estimated value is calculated to achieve sensor faults detection by comparing with the statistical threshold. What is more, the proposed faults detection approach is evaluated with two real flight datasets collected from the self-made rotorcraft and the experimental results show that the average ACC and AUC are increased by 3% and 2.3% respectively compared with the approach based on the standard Deep Forest.

An Intelligent Classification System for Cancer Detection Based on DNA Methylation Using ML and Semantic Knowledge in Healthcare.

To consistently assess a patient's internal and external wellness and diagnose chronic conditions like cancer, Alzheimer's disease, and cardiovascular disease, wearable sensing devices are being used. Wearable technologies and networking websites have become incredibly common in the medical sector in recent times. The condition of a patient's health can be influenced by a number of factors, including psychological response, emotional stability, and anxiety levels, which can be evaluated using social network analysis based on graph theory-based techniques and these ideas, known as "social network analysis" (SNA) are used to study relationship phenomena. Therefore, numerous uses for SNA in health research are possible, ranging from social science to exact science. For example, it can be used to research cooperative networks of healthcare providers and hazard-prone behaviors, infectious disease transmission, and the spread of initiatives for health promotion and prevention. Recently, a number of machine learning-based healthcare solutions have been proposed to track chronic illnesses utilizing data from social networks and wearable monitoring devices. In our suggested approach, we are using an intelligent system with the assistance of wearable sensors for the classification of cancer based on DNA methylation, an important epigenetic process in the human genome that controls gene expression and has been connected to a number of health issues. A mixed-sampling imbalanced data ensemble classification technique is created with the help of biomedical sensors to address the problem of class imbalance and high dimensionality in the Cancer Genome Atlas (TCGA) massive data. This technique is based on the Intelligent Synthetic Minority Oversampling (SMOTE) algorithm. The false-negative rate significantly rises as a result of this, to give a larger data set, a new minority class sample will be first obtained. The noise created during the sample expansion process is actually any data that has been acquired, preserved, or altered in a way that prevents the system that initially conceived it from accessing or utilizing it. Noisy data boosts the amount of space needed excessively and can also drastically influence the findings of any data collection investigation and therefore can also affect the sample sets of one or the other class, resulting in the class imbalance which acts as a common problem in ML datasets. The Tomek Link method is then used to eliminate this noise, producing a reasonably balanced data set. Each layer selects two random forest structures using the cascading forest structure of the deep forest (GC-Forest) algorithm to increase the generalization ability of the model and create the final classification model. Experiments using DNA methylation data collected by employing biosensors from six tumor patients reveal that the mixed-sampling unbalanced data ensemble classification technique may increase the sensitivity to the minority class while maintaining the majority class's classification accuracy.

An artificial intelligence diabetes management architecture based on 5G

Along with the development of 5G network and Internet of Things technologies, there has been an explosion in personalized healthcare systems. When the 5G and Artificial Intelligence (AI) is introduced into diabetes management architecture, it can increase the efficiency of existing systems and complications of diabetes can be handled more effectively by taking advantage of 5G. In this article, we propose a 5G-based Artificial Intelligence Diabetes Management architecture (AIDM), which can help physicians and patients to manage both acute complications and chronic complications. The AIDM contains five layers: the sensing layer, the transmission layer, the storage layer, the computing layer, and the application layer. We build a test bed for the transmission and application layers. Specifically, we apply a delay-aware RA optimization based on a double-queue model to improve access efficiency in smart hospital wards in the transmission layer. In application layer, we build a prediction model using a deep forest algorithm. Results on real-world data show that our AIDM can enhance the efficiency of diabetes management and improve the screening rate of diabetes as well.

DeepCPPred: A Deep Learning Framework for the Discrimination of Cell-Penetrating Peptides and Their Uptake Efficiencies.

Cell-penetrating peptides (CPPs) are special peptides capable of carrying a variety of bioactive molecules, such as genetic materials, short interfering RNAs and nanoparticles, into cells. Recently, research on CPP has gained substantial interest from researchers, and the biological mechanisms of CPPS have been assessed in the context of safe drug delivery agents and therapeutic applications. Correct identification and synthesis of CPPs using traditional biochemical methods is an extremely slow, expensive and laborious task particularly due to the large volume of unannotated peptide sequences accumulating in the World Bank repository. Hence, a powerful bioinformatics predictor that rapidly identifies CPPs with a high recognition rate is urgently needed. To date, numerous computational methods have been developed for CPP prediction. However, the available machine-learning (ML) tools are unable to distinguish both the CPPs and their uptake efficiencies. This study aimed to develop a two-layer deep learning framework named DeepCPPred to identify both CPPs in the first phase and peptide uptake efficiency in the second phase. The DeepCPPred predictor first uses four types of descriptors that cover evolutionary, energy estimation, reduced sequence and amino-acid contact information. Then, the extracted features are optimized through the elastic net algorithm and fed into a cascade deep forest algorithm to build the final CPP model. The proposed method achieved 99.45 percent overall accuracy with the CPP924 benchmark dataset in the first layer and 95.43 percent accuracy in the second layer with the CPPSite3 dataset using a 5-fold cross-validation test. Thus, our proposed bioinformatics tool surpassed all the existing state-of-the-art sequence-based CPP approaches.

Auto-tuning deep forest for shear stiffness prediction of headed stud connectors

The shear stiffness of headed stud connector is a critical parameter for the calculation of deflection and interfacial shear force for steel–concrete composite structure. Thus, this study presented a promising data-driven model auto-tuning Deep Forest (ATDF) to precisely predict the stud shear stiffness, where the novel Deep Forest algorithm is integrated with the Sequential Model-Based Optimization method to achieve automatic hyperparameter optimization. Six variables having causal relationships with shear stiffness were extracted via mechanism and model analysis, including the effect of weld collar that cannot be considered in existing models and subsequently constituting a database of 425 push-out tests. Then the ATDF model was trained by combining the advantages of deep learning, ensemble learning, and auto-tuning techniques. It was approved to significantly outperform representative benchmark models with R values of 0.91 and 0.87 for training and testing sets. The ATDF was subjected to attribute importance analysis, which quantified the stud diameter and concrete elastic modulus as the most significant variables for shear stiffness, with the stud elastic modulus having the minimal effect. The model uncertainty of ATDF was further evaluated, revealing that it had the lowest bias and variability than those in existing empirical or semi-empirical models. Finally, the reliability analysis was conducted and the partial factors of ATDF under specified target reliability were derived.

Integrating a deep forest algorithm with vector‐based cellular automata for urban land change simulation

AbstractWith the capacity to represent irregular geographical entities precisely, vector‐based cellular automata (VCA) have been extensively employed in urban land change simulation at the land parcel level. However, while more driving factors are considered, modeling the complicated nonlinear relationship between land parcel attributes and multiple land‐use changes is increasingly difficult. Moreover, in VCA, the driving factors are unstructured and cannot be directly modeled by traditional deep learning methods, which can only be applied to structured data processing. In order to address these problems, a new VCA model DF‐VCA that adopts the deep forest (DF) algorithm for mining CA transition rules was proposed to simulate urban land‐use change at the parcel level. The DF algorithm is a new deep learning method that can directly mine high‐level features from unstructured data and obtain accurate land‐use transition rules. The proposed DF‐VCA model was applied to simulate the urban land change in Shenzhen, China. Compared with several traditional VCA models, the DF‐VCA model achieved the best simulation performance at parcel‐level (Figure of Merit = 39.88%), pattern‐level (similarity = 96.47%), and community‐level (correlation coefficient = 0.9269). The results show that the DF‐VCA model with strong representative learning ability could simulate urban land change at fine scales precisely. Furthermore, the proposed DF‐VCA model was applied in Shenzhen's future land change simulations to guide sustainable urban development.

A novel performance degradation prognostics approach and its application on ball screw

The performance degradation prognostics of ball screw means important economic value and engineering application prospect. This paper proposes a performance degradation prognostics method which can be applied on ball screw. A clustering-based ensemble deep auto-encoders (EDAEs) was designed based on the selective ensemble and majority voting to extract features from the acceleration data. Then the sensitive feature distributions of different degradation cycles constitute the Gaussian mixture model (GMM), and the overlap degree of these distributions can be calculated to construct the health indicator. Finally, deep forest algorithm was introduced to achieve trend prognosis of health indicator. Meanwhile, the validity of the proposed method is confirmed by whole life cycle data of ball screw. The experimental results demonstrate that the proposed method can accurately identify the risk level of performance and realize the performance degradation prognostics.

A Fault Diagnosis Method of Rolling Bearing Based on Wavelet Packet Analysis and Deep Forest

The frequent accidents caused by the main fan motor in coal mines have exposed the safety hazards of rolling bearings. When a rolling bearing fails, its symmetry is broken, resulting in a rapid decline in its safety performance and posing a great threat to the main fan. Therefore, accurate rolling bearing fault diagnoses are the key to ensuring the safe and durable operation of main fans. Thus, in this paper, we propose a new fault diagnosis method of rolling bearing based on wavelet packet analysis and deep forest algorithm. Firstly, experiments were conducted under different health states to guarantee the diversity of data relating to the rolling bearing’s main fan and then to ensure the accuracy of the fault diagnosis under different health states. On the basis of the collected vibration signal data, we conducted the wavelet packet analysis method to extract the characteristics of the vibration signal and obtained a feature vector that characterizes the health of the bearing. After that, the extracted feature vector was used as the feature vector of the deep forest algorithm to train the deep forest diagnosis model and determine the location and fault type of the bearing fault. Finally, the proposed method in this paper was validated with real-time monitoring data of a main ventilation fan and compared with other diagnostic algorithms, which not only verified the diagnostic capability of deep forest in handling small samples, but also verified the diagnostic capability of the fault diagnosis model. In summary, the proposed fault diagnosis approach is promising in real coal mine main fans.

An Optimization Algorithm to Improve the Accuracy of Finger Vein Recognition

As people’s daily behavioral activities become more data-based, how to protect personal information security is a key consideration for the whole society. Finger vein recognition is becoming an important means of identification because of its uniqueness, live detection, security and many other advantages. Although deep learning can make finger vein recognition have good effect, but the number of samples needed to build deep network model is too large, and the current authoritative finger vein database cannot reach the minimum number of samples required. The emergence of Muti-Grained Cascade Forest provides a solution to the problem of insufficient sample data and long training time, which can give a new research avenue in feature extraction. In order to obtain higher accuracy, the deep forest algorithm is introduced in this paper to process the finger vein images. Firstly, the image data in the finger vein image database is pre-processed to prepare for the subsequent feature extraction and matching. Then, the deep forest algorithm is used to find the feature points and ORB algorithm is used to match the features to obtain the angular information of each matched pair, and the final identity is determined according to the sparse distribution of angles. The accuracy of finger vein recognition based on deep forest algorithm is 98.40%. By comparing with other machine learning methods for finger vein recognition, the proposed method has a higher accuracy rate.

A Novel Method for Predicting DNA N4-Methylcytosine Sites Based on Deep Forest Algorithm

A Novel Method for Predicting DNA N4-Methylcytosine Sites Based on Deep Forest Algorithm

Fault Diagnosis of Power Plant Condenser With the Optimized Deep Forest Algorithm

As an important component of power plant operation, condenser fault diagnosis plays a vital role in the safe and stable unit performance. However, the precision of most existing diagnostic methods is not high enough for condenser fault diagnosis. It is considerably difficult to diagnose a condenser fault even under various complicated conditions. In this study, a novel classification hybrid model (PCA-DF) combining the Principal Component Analysis (PCA) method with the Deep Forest (DF) model is proposed based on the motivation of improving the diagnosis accuracy of condenser fault. The algorithm of this hybrid model takes the dimension reduction result of the PCA method as the input to the DF model. The multigrained scanning structure and the dimension reduction method are considered to create a good effect. The experimental results verify the feasibility and effectiveness of this method on the historical fault sample data of the condenser. The focus on the work presented of this paper is to optimize the DF model based on PCA and study the fault diagnosis effect of the hybrid model. Results show that (1) the prediction accuracy for the condenser fault diagnosis can be improved by increasing the sample size with the PCA-DF method. (2) The accuracy of the results obtained by proposing the improved hybrid models is 1%-8% higher than the accuracy of the results obtained by directly introducing the DF model. The modified hybrid models still have advantages over the DF model for a small sample. (3) With an increase in the proportion of training sets, the accuracy of the modified hybrid models is improved correspondingly from 88.18% to 99.23%. (4) Compared with the backpropagation neural network, convolutional neural network, relevance vector machine and kernel Fisher discriminant analysis models, the PCA-DF model has higher accuracy. In this study, the proposed models can eliminate the influence of autocorrelation between data, and condenser fault diagnosis based on modified models has the fastest convergence speed and best accuracy. Furthermore, the proposed novel models can be extended to more complex fault diagnosis in other fields.

Deep Forest classifier for wetland mapping using the combination of Sentinel-1 and Sentinel-2 data

ABSTRACT Wetlands are among the most important, yet in danger ecosystems and play a vital role for the well-being of humans as well as flora and fauna. Over the past few years, state-of-the-art deep learning (DL) tools have gained attention for wetland classification within the remote sensing community. However, the DL methods could have complex structure and their efficiency greatly depends on the availability of a large number of training data. Inspired by DL methods, yet with less complexity, the Deep Forest (DF) classifier is an advanced tree-based deep learning tool with a great capability for several remote sensing applications. Despite the effectiveness of DF classifiers, few research studies have investigated the potential of such a powerful technique for classification of remote sensing, with no documented research for wetland classification. Accordingly, the potential of the DF algorithm for the classification of wetland complexes has been investigated in this study. In particular, three well-known classifiers, namely Extreme Gradient Boosting (XGB), Random Forest (RF), and Extra Tree (ET), were used as the tree-based classifier to build DF, for which the hyper parameter tuning is carried out to ensure the optimum classification accuracy. Three well-known tree-based classification algorithms, namely Decision Tree (DT), Conventional Random Forest (CRF), and Conventional Extreme Gradient Boosting (CXGB), as well as a Convolutional Neural Network (CNN) are used as benchmark tools to compare the results obtained from the DF classifiers for wetland mapping. The results demonstrated that the DF-XGB classifier outperforms both DF-RF and DF-ET in terms of classification accuracy albeit with a longer training time. The results also confirmed the superiority of all three DF-based classifiers compared to the CRF and DT classifiers. For example, the DF-XGB improved the F1-score by 14%, 13%, 7%, 3%, and 1% for fen, swamp, marsh, bog, and shallow water, respectively, compared to the optimized CRF. The results indicated that the DF algorithm has great capability to be applied over large areas to support regional and national wetland mapping and monitoring.

Variable length particle swarm optimization and multi-feature deep fusion for motor imagery EEG classification

Brain-computer interfaces are a new pathway for communication between human body and the external environment. High classification accuracy for motor imagery electroencephalogram (EEG) signals is desirable by improving the algorithm of feature extraction and classification. A novel algorithm (VLPSO-MFDF) based on the variable length particle swarm optimization (VLPSO) and multi-feature deep fusion (MFDF) is proposed. First, each layer of the deep forest is reconstructed into two same classification modules. Then, several different features are extracted for the motor imagery EEG signal to feed separately to the classification modules. The VLPSO is used to search for the optimal weights for the probability vectors output by each classification module, which can continuously optimize the classification performance. Experimental results demonstrate that the VLPSO-MFDF algorithm can achieve higher classification accuracy for four classifications of motor imagery EEG signals compared with the traditional deep forest algorithm. The proposed method fused multi-domain features and corrected the prediction difference. It was of great significance for improving the performance of the classifier.

The Fault Diagnosis of Rolling Bearing Based on Improved Deep Forest

Rolling bearing fault diagnosis is a meaningful and challenging task. Most methods first extract statistical features and then carry out fault diagnosis. At present, the technology of intelligent identification of bearing mostly relies on deep neural network, which has high requirements for computer equipment and great effort in hyperparameter tuning. To address these issues, a rolling bearing fault diagnosis method based on the improved deep forest algorithm is proposed. Firstly, the fault feature information of rolling bearing is extracted through multigrained scanning, and then the fault diagnosis is carried out by cascade forest. Considering the fitting quality and diversity of the classifier, the classifier and the cascade strategy are updated. In order to verify the effectiveness of the proposed method, a comparison is made with the traditional machine learning method. The results suggest that the proposed method can identify different types of faults more accurately and robustly. At the same time, it has very few hyperparameters and very low requirements on computer hardware.

A real-time deep learning forest fire monitoring algorithm based on an improved Pruned + KD model

To meet the needs of embedded intelligent forest fire monitoring systems using an unmanned aerial vehicles (UAV), a deep learning fire recognition algorithm based on model compression and lightweight requirements is proposed in this study. The algorithm for the lightweight MobileNetV3 model was developed to reduce the complexity of the conventional YOLOv4 network structure. The redundant channels are eliminated through channel-level sparsity-induced regularization. The knowledge distillation algorithm is used to improve the detection accuracy of the pruned model. The experimental results reveal that the number of model parameters for the proposed architecture is only 2.64 million—compared with YOLOv4, this represents a reduction of nearly 95.87%. The inference time decreased from 153.8 to 37.4 ms, a reduction of nearly 75.68%. Our approach shows the advantages of a model with a smaller number of parameters, low memory requirements and fast inference speed compared with existing algorithms. The method presented in this paper is specifically tailored for use as a deep learning forest fire monitoring system on a UAV platform.