Nonlinear Classification Research Articles

Predicting the Bioconcentration Factor in Fish from Molecular Structures.

The bioconcentration factor (BCF) is one of the metrics used to evaluate the potential of a substance to bioaccumulate into aquatic organisms. In this work, linear and non-linear regression QSARs were developed for the prediction of log BCF using different computational approaches, and starting from a large and structurally heterogeneous dataset. The new MLR-OLS and ANN regression models have good fitting with R2 values of 0.62 and 0.70, respectively, and comparable external predictivity with R2ext 0.64 and 0.65 (RMSEext of 0.78 and 0.76), respectively. Furthermore, linear and non-linear classification models were developed using the regulatory threshold BCF >2000. A class balanced subset was used to develop classification models which were applied to chemicals not used to create the QSARs. These classification models are characterized by external and internal accuracy up to 84% and 90%, respectively, and sensitivity and specificity up to 90% and 80%, respectively. QSARs presented in this work are validated according to regulatory requirements and their quality is in line with other tools available for the same endpoint and dataset, with the advantage of low complexity and easy application through the software QSAR-ME Profiler. These QSARs can be used as alternatives for, or in combination with, existing models to support bioaccumulation assessment procedures.

CT-based transformer model for non-invasively predicting the Fuhrman nuclear grade of clear cell renal cell carcinoma.

BackgroundClear cell Renal Cell Carcinoma (ccRCC) is the most common malignant tumor in the urinary system and the predominant subtype of malignant renal tumors with high mortality. Biopsy is the main examination to determine ccRCC grade, but it can lead to unavoidable complications and sampling bias. Therefore, non-invasive technology (e.g., CT examination) for ccRCC grading is attracting more and more attention. However, noise labels on CT images containing multiple grades but only one label make prediction difficult. However, noise labels exist in CT images, which contain multiple grades but only one label, making prediction difficult.AimWe proposed a Transformer-based deep learning algorithm with CT images to improve the diagnostic accuracy of grading prediction and to improve the diagnostic accuracy of ccRCC grading.MethodsWe integrate different training models to improve robustness and predict Fuhrman nuclear grade. Then, we conducted experiments on a collected ccRCC dataset containing 759 patients and used average classification accuracy, sensitivity, specificity, and AreaUnderCurve as indicators to evaluate the quality of research. In the comparative experiments, we further performed various current deep learning algorithms to show the advantages of the proposed method. We collected patients with pathologically proven ccRCC diagnosed from April 2010 to December 2018 as the training and internal test dataset, containing 759 patients. We propose a transformer-based network architecture that efficiently employs convolutional neural networks (CNNs) and self-attention mechanisms to extract a persuasive feature automatically. And then, a nonlinear classifier is applied to classify. We integrate different training models to improve the accuracy and robustness of the model. The average classification accuracy, sensitivity, specificity, and area under curve are used as indicators to evaluate the quality of a model.ResultsThe mean accuracy, sensitivity, specificity, and Area Under Curve achieved by CNN were 82.3%, 89.4%, 83.2%, and 85.7%, respectively. In contrast, the proposed Transformer-based model obtains a mean accuracy of 87.1% with a sensitivity of 91.3%, a specificity of 85.3%, and an Area Under Curve (AUC) of 90.3%. The integrated model acquires a better performance (86.5% ACC and an AUC of 91.2%).ConclusionA transformer-based network performs better than traditional deep learning algorithms in terms of the accuracy of ccRCC prediction. Meanwhile, the transformer has a certain advantage in dealing with noise labels existing in CT images of ccRCC. This method is promising to be applied to other medical tasks (e.g., the grade of neurogliomas and meningiomas).

ECNN: Intelligent Fault Diagnosis Method Using Efficient Convolutional Neural Network

With outstanding deep feature learning and nonlinear classification abilities, Convolutional Neural Networks (CNN) have been gradually applied to deal with various fault diagnosis tasks. Affected by variable working conditions and strong noises, the empirical datum always has different probability distributions, and then different data segments may have inconsistent contributions, so more attention should be assigned to the informative data segments. However, most of the CNN-based fault diagnosis methods still retain black-box characteristics, especially the lack of attention mechanisms and ignoring the special contributions of informative data segments. To address these problems, we propose a new intelligent fault diagnosis method comprised of an improved CNN model named Efficient Convolutional Neural Network (ECNN). The extensive view can cover the special characteristic periods, and the small view can locate the essential feature using Pyramidal Dilated Convolution (PDC). Consequently, the receptive field of the model can be greatly enlarged to capture the location information and excavate the remarkable informative data segments. Then, a novel residual network feature calibration and fusion (ResNet-FCF) block was designed, which uses local channel interactions and residual networks based on global channel interactions for weight-redistribution. Therefore, the corresponding channel weight is increased, which puts more attention on the information data segment. The ECNN model has achieved encouraging results in information extraction and feature channel allocation of the feature. Three experiments are used to test different diagnosis methods. The ECNN model achieves the highest average accuracy of fault diagnosis. The comparison results show that ECNN has strong domain adaptation ability, high stability, and superior diagnostic performance.

Precision Medicine Approaches with Metabolomics and Artificial Intelligence.

Recent technological innovations in the field of mass spectrometry have supported the use of metabolomics analysis for precision medicine. This growth has been allowed also by the application of algorithms to data analysis, including multivariate and machine learning methods, which are fundamental to managing large number of variables and samples. In the present review, we reported and discussed the application of artificial intelligence (AI) strategies for metabolomics data analysis. Particularly, we focused on widely used non-linear machine learning classifiers, such as ANN, random forest, and support vector machine (SVM) algorithms. A discussion of recent studies and research focused on disease classification, biomarker identification and early diagnosis is presented. Challenges in the implementation of metabolomics–AI systems, limitations thereof and recent tools were also discussed.

Differential Privacy Protection for Support Vector Machines for Nonlinear Classification

Currently, private data leakage and nonlinear classification are two challenges encountered in big data mining. In particular, few studies focus on these issues in support vector machines (SVMs). In this paper, to effectively solve them, we propose a novel framework based on the concepts of differential privacy (DP) and kernel functions. This framework can allocate privacy budgets and add artificial noise to different SVM locations simultaneously, which makes the perturbation process freer and more delicate. In addition, under this framework, we propose three algorithms, DP SVMs that perturb the training data set, perturb the kernel function, and utilize mixed perturbation (DPSVM-TDP, DPSVM-KFP, and DPSVM-MP, respectively), all of which can realize accurate classification while ensuring that the users’ privacy is not violated. Moreover, we conduct privacy analysis on these algorithms and prove that they all satisfy ε , 0 − DP. Finally, we conduct experiments to evaluate the algorithms in terms of different aspects and compare them with the DPSVM with dual-variable perturbation (DVP) algorithm (DPSVM-DVP) to determine the optimal perturbation method. The results show that DPSVM-KFP can achieve the highest data utility and strictest privacy protection with the shortest running time.

Heart Disease Prediction Using Genetic Algorithm with Machine Learning Classifiers

Based on the world health organization (WHO), heart disease is the major reason for the loss of life everywhere on earth. Heart attacks are the leading reason of loss of life among heart diseases. This disease is called a silent disease. The person does not feel any pain until the last level of sickness and may arrive at death if not saved at the right time. The datasets for this disease are becoming available, so it is a particularly good branch of study. Predicting a heart attack for a medical practitioner is difficult since it requires increased expertise. However, over the last few decades, resolving complicated, extremely non-linear classification and prediction problems have been using machine learning algorithms (ML). Hence, it is feasible to establish a prediction model that would see the existence or nonexistence of heart disease based on many heart-related symptoms (features). The essential contribution of this research is to introduce various prediction models for heart disease using a genetic algorithm (GA) to find optimal features combined with classical machine learning techniques. The optimized prediction model uses a genetic algorithm that performs better than classical models. The execution of the algorithms is tested using Cleveland and Framingham datasets. The prediction models' performance is standardized using three evaluation criteria: accuracy, precision, and recall. The proposed system showed superior performance compared with other related systems. It reached an accuracy of 100% for the Cleveland dataset and 91.8% for the Framingham dataset.

Pattern classification based on regional models

In a supervised setting, the global classification paradigm leverages the whole training data to produce a single class discriminative model. Alternatively, the local classification approach builds multiple base classifiers, each of them using a small subset of the training data. In this paper, we take a path to stand in-between the global and local approaches. We introduce a two-level clustering-based method in which base classifiers operate on a larger portion of the input space than in the traditional local paradigm. In particular, we first obtain a grained input representation by employing a Self-Organizing Map (SOM) to the inputs. We then apply a clustering algorithm (e.g., K-Means) to the SOM units to define input regions — a subset of input samples associated with a specific cluster of SOM units. We refer to this approach as regional classification. We demonstrate the effectiveness of regional classification on several benchmarks. Also, we study the impact of (1) adopting linear and nonlinear base classifiers (e.g., least squares support vector machines) and (2) using cluster validation indexes to determine the optimal number of clusters. Based on the experiments, the regional classification approach achieves competitive performance compared to its global and local counterparts, especially when equipped with linear base classifiers.

Hessian Semisupervised Scatter Regularized Classification Model With Geometric and Discriminative Information for Nonlinear Process.

The integration of semisupervised modeling and discriminative information has been sporadically discussed in the research literature of traditional classification modeling, while the former one would make full use of the collected data and the latter one would further improve the classification performance. In this article, the Hessian semisupervised scatter regularized classification model is proposed as a coherent framework for the nonlinear process classification upon both labeled and unlabeled data. It is innovatively designed with a loss function to evaluate the classification accuracy and three regularization terms, respectively, corresponding to the geometry information, discriminative information, and model complexity. Both cases of the coherent framework, respectively, casted to the reproducing kernel Hilbert space and linear space, enjoy a theoretically guaranteed analytical solution. Experiments on process classification tasks on a benchmark dataset and a real industrial polyethylene process illustrate the merits of the proposed method in a sense that the class information of novel collected data is accurately predicted.

Exploiting Read Current Noise of TiOx Resistive Memory by Controlling Forming Conditions for Probabilistic Neural Network Hardware

Conductance variations of resistive random-access memory (RRAM) are significant challenges that hinder the accurate inference of neural network (NN) hardware. In this study, we exploit the read noise of the RRAM as an active computational enabler for implementing probabilistic NN. As electrical characteristics of RRAM are directly related to the properties of conductive filament (CF), we statistically explore read current of TiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</sub> -based RRAM with different forming conditions and explain the results by linking the CF model. In addition, an array mapping scheme to transfer weights to one transistor-one RRAM (1T1R) array is experimentally demonstrated. Through NN simulations, we verify that the probabilistic NN shows promising results on nonlinear classification problem avoiding overconfidence compared with deterministic NN.

Classification of cardiac electrical signals between patients with myocardial infarction and normal subjects by using nonlinear dynamics features and different classification models

Myocardial infarction (MI) is one of the leading causes of human mortality and morbidity around the world. Despite that much progress has been made for MI detection based on medical image analysis in recent years, most of them suffer from their expensiveness and invasive nature. In this paper, we demonstrate abnormalities of electrocardiogram (ECG) in a new quantifiable manner using nonlinear dynamics features and different classification models. Time-varying ECG data are represented as three-dimensional vectorcardiogram (VCG) and the underlying cardiac dynamics. Nonlinear dynamics measures, including entropy variability measures, complexity measures and chaotic measures, are calculated and fed into different machine learning methods for the classification task. The extracted nonlinear dynamics measures reflect in-depth cardiac dynamics characteristic, which is shown to be more sensitive to subtle ECG modifications. Therefore, it is expected to provide an early detection tool for latent ECG modifications before obvious diagnostic changes are observed. Experiments on the PTB database are conducted to demonstrate the efficiency of the proposed method.

Linear optical sampling enabled soliton nonlinear frequency spectrum classification.

Nonlinear Fourier transform (NFT) is a powerful tool for characterizing optical soliton dynamics, which, however, suffers from fundamental limitations that ultra-wide bandwidth photodetectors and ultra-high sampling rate analog-to-digital converters should be used when accessing the full-field information of an ultrafast optical pulse. Herein, we report on the experimental demonstration of the linear optical sampling (LOS) enabled nonlinear frequency spectrum classification of ultrashort optical pulses, which could break this limitation. Instead of traditional coherent detection, the LOS overcomes the ultra-wide bandwidth constraint of commercially available optoelectrical devices. By finely adjusting the repetition rate difference between the soliton to be characterized and the sampling pulsed source, a 55.56-TSa/s equivalent sampling rate arising in the LOS can be secured, where only 400-MHz balanced photodetectors and 5-GSa/s analog-to-digital converter are used. Meanwhile, according to the nonlinear frequency spectrum calculated from the accurate full-field information, the promising concept of soliton distillation has been experimentally verified for the first time. The LOS-enabled NFT technique provides an alternative and efficient characterization tool for ultrafast fiber lasers, which facilities comprehensive insight into soliton dynamics.

Automatic epileptic seizure detection in EEG signals using sparse common spatial pattern and adaptive short-time Fourier transform-based synchrosqueezing transform

Epilepsy can now be diagnosed more accurately and quickly due to computer-aided seizure detection utilizing Electroencephalography (EEG) recordings. In this work, a novel method for automated seizure identification from the EEG signal is proposed utilizing the sparse common spatial pattern (sCSP) and the adaptive short-time Fourier transform-based synchrosqueezing transform (adaptive FSST). The sCSP is utilized to select optimal channels and discriminate seizure states. In order to enhance the time–frequency representation of multi-component EEG signals and reduce noise and interferences, the selected channels are imported into adaptive FSST. The frequency spectrum of the adaptive FSST is separated into distinct segments concentrated on a particular frequency to improve seizure detection by adapting the EEG signal decomposition to the rhythmic component of seizures. Thus, a simple linear classifier is all that is needed to identify seizures thanks to this collection of algorithms since the extracted features from each segment’s instantaneous amplitude and frequency and logvariance of each selected channel could be highly discriminative. Using linear classifiers for these discriminative features avoids overfitting and training complexity inherent in most non-linear classifiers and deep learning approaches. The proposed approach is compared to previous high-performance algorithms using the CHB-MIT epileptic EEG database. This method outperformed most of the best findings in the literature regarding mean performance, with mean sensitivity, specificity, and accuracy of 98.44%, 99.19%, and 98.81%, respectively. Thus, the proposed method’s outstanding performance demonstrates that it could significantly decrease EEG interpretation load on specialists and be a powerful tool for an epilepsy diagnosis.

Machine Learning-Based 30-Day Hospital Readmission Predictions for COPD Patients Using Physical Activity Data of Daily Living with Accelerometer-Based Device.

Chronic obstructive pulmonary disease (COPD) is a significantly concerning disease, and is ranked highest in terms of 30-day hospital readmission. Generally, physical activity (PA) of daily living reflects the health status and is proposed as a strong indicator of 30-day hospital readmission for patients with COPD. This study attempted to predict 30-day hospital readmission by analyzing continuous PA data using machine learning (ML) methods. Data were collected from 16 patients with COPD over 3877 days, and clinical information extracted from the patients’ hospital records. Activity-based parameters were conceptualized and evaluated, and ML models were trained and validated to retrospectively analyze the PA data, identify the nonlinear classification characteristics of different risk factors, and predict hospital readmissions. Overall, this study predicted 30-day hospital readmission and prediction performance is summarized as two distinct approaches: prediction-based performance and event-based performance. In a prediction-based performance analysis, readmissions predicted with 70.35% accuracy; and in an event-based performance analysis, the total 30-day readmissions were predicted with a precision of 72.73%. PA data reflect the health status; thus, PA data can be used to predict hospital readmissions. Predicting readmissions will improve patient care, reduce the burden of medical costs burden, and can assist in staging suitable interventions, such as promoting PA, alternate treatment plans, or changes in lifestyle to prevent readmissions.

A Human Motion Function Rehabilitation Monitoring System Based on Data Mining

Human motion interaction technologies have evolved to a new level with the development of traditional reality technology as science and technology have developed. Fully interactive and human motion interaction technologies are becoming more common in fields such as medical rehabilitation and military simulations. Human motion is at the heart of all activity, and motion analysis and human motion are critical theoretical disciplines. Identification is based on behavior and motion in human motion, with attributes such as effectiveness, intelligence, potent interaction, and rich expression data. When studying human movement, many researchers now prefer this method. However, this study was conducted with insufficient suggestions for real-time human motion function assessment, rehabilitation, and improvement. The development of an information monitoring system for human motion function rehabilitation can be used to evaluate the efficacy of patient rehabilitation training. A human motion function rehabilitation monitoring system is created using an effective and thorough design methodology. The system is made up of the rehabilitation monitoring terminal, the human motion function monitoring module, and the medical center monitoring system. Therefore, the motion-based data mining technique is better for the human motion function rehabilitation monitoring system. The normalized proportion of motion features will assist in the creation of a database for human motion mining. The nonlinear classification function is used in this paper to scientifically categorize human motion features to implement data mining techniques for monitoring human motion function rehabilitation. The effectiveness of patient rehabilitation is significantly increased by the use of a human motion-based rehabilitation monitoring system.

Common, uncommon, and novel applications of random forest in psychological research.

Recent reform efforts have pushed toward a better understanding of the distinction between exploratory and confirmatory research, and appropriate use of each. As some utilize more exploratory tools, it may be tempting to employ multiple linear regression models. In this paper, we advocate for the use of random forest (RF) models. RF is able to obtain better predictive performance than traditional regression, while also inherently protecting against overfitting as well as detecting nonlinear effects and interactions among predictors. Given the advantages of RF compared to other statistical procedures, it is a tool commonly used within a plethora of industries, including stock trading, banking, pharmaceuticals, and patient healthcare planning. However, we find RF is used within the field of psychology comparatively less frequently. In the current paper, we advocate for RF as an important statistical tool within the context of behavioral and psychological research. In hopes of increasing the use of RF in the field of psychology, we provide information pertaining to the limitations one might confront in using RF and how to overcome such limitations. Moreover, we discuss various methods for how to optimally utilize RF with psychological data, such as nonparametric modeling, interaction and nonlinearity detection, variable selection, prediction and classification modeling, and assessing parameters of Monte Carlo simulations. Throughout, we illustrate the use of RF with visualization strategies, aimed to make RF models more comprehensible and intuitive.

A Spectral Feature Selection Approach With Kernelized Fuzzy Rough Sets

Feature evaluation is an important issue in constructing a feature selection algorithm in kernelized fuzzy rough sets, which has been proven to be an effective approach to deal with nonlinear classification tasks and uncertainty in learning problems. However, the feature evaluation function developed with kernelized fuzzy rough sets cannot better reflect the affinity relationship of samples and is time-consuming. To overcome these drawbacks, in this article, the problem of feature selection with kernelized fuzzy rough sets is studied based on the spectral graph theory. First, the within-class and between-class sample similarity matrices by using kernelized fuzzy approximation operators are constructed. Two operators, which can capture the affinity relationship of samples, are then introduced based on the sample similarity matrices. The proposed operator can be regarded as the sum of the weighted kernelized fuzzy approximation operators. Second, based on the ratio criterion, a feature evaluation function and its corresponding feature selection algorithm FRKF are presented, which can effectively evaluate the importance of features. Third, to illustrate the performance of the proposed algorithm, extensive experiments have been carried out to compare FRKF and other well-known feature selection methods, including the feature ranking methods and feature subset selection methods on various classification tasks. The experimental results on real-world datasets demonstrate that FRKF achieves the high performances in terms of the robustness, efficiency, and effectiveness.

Support Vector Machine with Robust Low-Rank Learning for Multi-Label Classification Problems in the Steelmaking Process

In this paper, we present a novel support vector machine learning method for multi-label classification in the steelmaking process. The steelmaking process involves complicated physicochemical reactions. The end-point temperature is the key to the steelmaking process. According to the initial furnace condition information, the end-point temperature can be predicted using a data-driven method. Based on the setting value of the temperature before tapping, multi-scale predicted errors of the end-point temperature can be calculated and divided into different ranges. The quality evaluation problem can be attributed to the multi-label classification problem of molten steel quality. To solve the classification problem, considering that it is difficult to capture nonlinear relationships between the input and output in linear models, we propose a novel support vector machine with robust low-rank learning, which has the characteristics of class imbalance without label correlations; a low-rank constraint is used to deal with high-order label correlations in low-dimensional space. Furthermore, we derive an accelerated proximal gradient algorithm and then extend it to handle the nonlinear multi-label classifiers. To validate the proposed model, experiments are conducted with real data from a practical steelmaking problem. The results show that the proposed model can effectively solve the multi-label classification problem in industrial production. To evaluate the proposed approach as a general classification approach, we test it on multi-label classification benchmark datasets. The results illustrate that the proposed approach performs better than other state-of-the-art approaches across different scenarios.

A kernel-free fuzzy reduced quadratic surface [formula omitted]-support vector machine with applications

The kernel-free support vector machine (SVM) models are recently developed and studied to overcome some drawbacks induced by the kernel-based SVM models. To further improve the classification accuracy and computational efficiency of existing kernel-free quadratic surface support vector machine (QSSVM) models, a novel kernel-free ν-fuzzy reduced QSSVM model is proposed. The proposed model utilizes a reduced quadratic surface for nonlinear binary classification as well as reducing the effect of outliers in the data set. Some theoretical properties are rigorously studied, especially, the effects of the parameter ν on the dual feasibility and the number of support vectors. Computational experiments are conducted on some public benchmark data sets to indicate the superior performance of the proposed model over some well-known binary classification models. The numerical results also favors the higher training efficiency of the proposed model over those of other kernel-free SVM models. Moreover, the proposed model is successfully applied to the prodromal detection of Alzheimer’s Disease with good performance, by using the data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database.

Simplified Transfer Learning for Chest Radiography Models Using Less Data

Background Developing deep learning models for radiology requires large data sets and substantial computational resources. Data set size limitations can be further exacerbated by distribution shifts, such as rapid changes in patient populations and standard of care during the COVID-19 pandemic. A common partial mitigation is transfer learning by pretraining a "generic network" on a large nonmedical data set and then fine-tuning on a task-specific radiology data set. Purpose To reduce data set size requirements for chest radiography deep learning models by using an advanced machine learning approach (supervised contrastive [SupCon] learning) to generate chest radiography networks. Materials and Methods SupCon helped generate chest radiography networks from 821 544 chest radiographs from India and the United States. The chest radiography networks were used as a starting point for further machine learning model development for 10 prediction tasks (eg, airspace opacity, fracture, tuberculosis, and COVID-19 outcomes) by using five data sets comprising 684 955 chest radiographs from India, the United States, and China. Three model development setups were tested (linear classifier, nonlinear classifier, and fine-tuning the full network) with different data set sizes from eight to 85. Results Across a majority of tasks, compared with transfer learning from a nonmedical data set, SupCon reduced label requirements up to 688-fold and improved the area under the receiver operating characteristic curve (AUC) at matching data set sizes. At the extreme low-data regimen, training small nonlinear models by using only 45 chest radiographs yielded an AUC of 0.95 (noninferior to radiologist performance) in classifying microbiology-confirmed tuberculosis in external validation. At a more moderate data regimen, training small nonlinear models by using only 528 chest radiographs yielded an AUC of 0.75 in predicting severe COVID-19 outcomes. Conclusion Supervised contrastive learning enabled performance comparable to state-of-the-art deep learning models in multiple clinical tasks by using as few as 45 images and is a promising method for predictive modeling with use of small data sets and for predicting outcomes in shifting patient populations. © RSNA, 2022 Online supplemental material is available for this article.

Prediction of corporate financial distress based on digital signal processing and multiple regression analysis

Abstract In order to reduce the default rate of corporate bond market, the author proposes to use digital signal processing and multiple regression analysis to study the prediction system of financial distressed companies. First, design the research method, Logistic regression model is the most commonly used multivariate statistical method when modeling binary dependent variables, it can solve the problem of nonlinear classification, it has no specific requirements for the distribution of variables, and the accuracy of judgment is high. The author selects 32 financial ratios from the perspectives of solvency, operating ability, profitability, development ability, per share index, and risk level. Taking special treatment (ST) due to abnormal financial status as a sign of financial distress in listed companies, when selecting samples, the matching principle is adopted to select non-ST companies as matching samples. Two methods of logistic regression and support vector machine are used for empirical testing, and both in-sample testing and out-of-sample prediction are performed. The results show that when using the logistic regression method, the propensity to default indicator (TTD) reflected in the text content, it can indeed improve the out-of-sample prediction accuracy of the financial distress prediction model, and it is consistent with the in-sample test, this is mainly reflected in the reduction of the first type of error, that is, the probability of misjudging a financially distressed company as a normal company. Changes in the proportions have little effect on the relative importance of financial ratio variables when modeling with support vector machines, the propensity to default indicator (TTD) entered the top ten important variables in both ratios, and ranked fourth among all indicators when the ratio was 1:2, importance has increased significantly. From this it can be seen that, when using support vector machine to build a financial distress prediction model, the propensity to default indicator (TTD) has played an important role. In the case of using the support vector machine method, adding the default tendency indicator (TTD) reflected by the text information can also improve the accuracy of the financial distress prediction model.