Probabilistic Classification Model Research Articles

A novel space–spectrum array tile probability random-forest model enhances LULC mapping accuracy on Google Earth Engine: An experiment in Ordos, China

The rapid renewal of land use and land cover (LULC) maps using remote-sensing technologies constitutes a sine qua non for judicious land resource management at both regional and national scales. Existing research conducted on the Google Earth Engine (GEE) platform has overwhelmingly focused on pixel-based LULC classification techniques, often neglecting the role of spatial context via neighbouring valuable pixel information. Remarkably, little attention has been paid to the amalgamation of 3 × 3 neighbouring pixels into a three-dimensional space–spectrum array that can emulate the functionalities of object-based image analysis. In this study, we developed a novel integrated model consisting of a space–spectrum array (SSA) model based on 3 × 3 neighbouring pixels, a tile model based on random forest, and a multiple probabilistic classification model (SSA-TPRF) on the GEE platform to generate a LULC map with high overall accuracy (OA) for Ordos in 2020. Three bimonthly median value images were synthesised and feature collections, including spectral bands and vegetation indices, were constructed. Five experimental groups (EXP1–EXP5) were used to assess the different model combinations. Subsequent validation procedures employed abundant reference samples and compared the results with those of the three extant LULC mapping products. The results showed that EXP2, which was grounded in the tile-based model, yielded an OA of 87.53%, surpassing that of EXP1 (84.99%), which employed a traditional overall model. Furthermore, EXP3, which integrated the multiple probabilistic classification model with the traditional overall model, exhibited an OA of 85.19%, exceeding that of EXP1. A comparison of the five experimental groups using the four regional spatial subtlety features revealed that the EXP5, employing the SSA-TPRF model, successfully decreased the salt and pepper noise. The OA of six tile sizes ranging from 10 km to 100 km were compared, and the highest OA (88.35%) was achieved at a tile size of 25 km. The resultant LULC map in Ordos, derived from the SSA-TPRF model, showed superior OA compared with the extant LULC products. This study thus contributes to a versatile and scalable model within the GEE framework, offering avenues for facile adaptation and recurrent application across disparate geographical locations and temporal settings. The adaptability of this model is particularly advantageous for developing nations and regions typified by diverse landscapes, thereby catalysing the iterative updating of LULC maps through advanced remote-sensing paradigms.

Learning to Learn in Interactive Constraint Acquisition

Constraint Programming (CP) has been successfully used to model and solve complex combinatorial problems. However, modeling is often not trivial and requires expertise, which is a bottleneck to wider adoption. In Constraint Acquisition (CA), the goal is to assist the user by automatically learning the model. In (inter)active CA, this is done by interactively posting queries to the user, e.g. does this partial solution satisfy your (unspecified) constraints or not. While interactive CA methods learn the constraints, the learning is related to symbolic concept learning, as the goal is to learn an exact representation. However, a large number of queries is required to learn the model, which is a major limitation. In this paper, we aim to alleviate this limitation by tightening the connection of CA and Machine Learning (ML), by, for the first time in interactive CA, exploiting statistical ML methods. We propose to use probabilistic classification models to guide interactive CA queries to the most promising parts. We discuss how to train classifiers to predict whether a candidate expression from the bias is a constraint of the problem or not, using both relation-based and scope-based features. We then show how the predictions can be used in all layers of interactive CA: the query generation, the scope finding, and the lowest-level constraint finding. We experimentally evaluate our proposed methods using different classifiers and show that our methods greatly outperform the state of the art, decreasing the number of queries needed to converge by up to 72%.

Development of an individualized risk calculator of treatment resistance in patients with first-episode psychosis (TRipCal) using automated machine learning: a 12-year follow-up study with clozapine prescription as a proxy indicator.

About 15-40% of patients with schizophrenia are treatment resistance (TR) and require clozapine. Identifying individuals who have higher risk of development of TR early in the course of illness is important to provide personalized intervention. A total of 1400 patients with FEP enrolled in the early intervention for psychosis service or receiving the standard psychiatric service between July 1, 1998, and June 30, 2003, for the first time were included. Clozapine prescriptions until June 2015, as a proxy of TR, were obtained. Premorbid information, baseline characteristics, and monthly clinical information were retrieved systematically from the electronic clinical management system (CMS). Training and testing samples were established with random subsampling. An automated machine learning (autoML) approach was used to optimize the ML algorithm and hyperparameters selection to establish four probabilistic classification models (baseline, 12-month, 24-month, and 36-month information) of TR development. This study found 191 FEP patients (13.7%) who had ever been prescribed clozapine over the follow-up periods. The ML pipelines identified with autoML had an area under the receiver operating characteristic curve ranging from 0.676 (baseline information) to 0.774 (36-month information) in predicting future TR. Features of baseline information, including schizophrenia diagnosis and age of onset, and longitudinal clinical information including symptoms variability, relapse, and use of antipsychotics and anticholinergic medications were important predictors and were included in the risk calculator. The risk calculator for future TR development in FEP patients (TRipCal) developed in this study could support the continuous development of data-driven clinical tools to assist personalized interventions to prevent or postpone TR development in the early course of illness and reduce delay in clozapine initiation.

Relating the One-Parameter Logistic Diagnostic Classification Model to the Rasch Model and One-Parameter Logistic Mixed, Partial, and Probabilistic Membership Diagnostic Classification Models

Diagnostic classification models (DCMs) are statistical models with discrete latent variables (so-called skills) to analyze multiple binary variables (i.e., items). The one-parameter logistic diagnostic classification model (1PLDCM) is a DCM with one skill and shares desirable measurement properties with the Rasch model. This article shows that the 1PLDCM is indeed a latent class Rasch model. Furthermore, the relationship of the 1PLDCM to extensions of the DCM to mixed, partial, and probabilistic memberships is treated. It is argued that the partial and probabilistic membership models are also equivalent to the Rasch model. The fit of the different models was empirically investigated using six datasets. It turned out for these datasets that the 1PLDCM always had a worse fit than the Rasch model and mixed and partial membership extensions of the DCM.

UAV-Based Disease Detection in Palm Groves of Phoenix canariensis Using Machine Learning and Multispectral Imagery

Climate change and the appearance of pests and pathogens are leading to the disappearance of palm groves of Phoenix canariensis in the Canary Islands. Traditional pathology diagnostic techniques are resource-demanding and poorly reproducible, and it is necessary to develop new monitoring methodologies. This study presents a tool to identify individuals infected by Serenomyces phoenicis and Phoenicococcus marlatti using UAV-derived multispectral images and machine learning. In the first step, image segmentation and classification techniques allowed us to calculate a relative prevalence of affected leaves at an individual scale for each palm tree, so that we could finally use this information with labelled in situ data to build a probabilistic classification model to detect infected specimens. Both the pixel classification performance and the model’s fitness were evaluated using different metrics such as omission and commission errors, accuracy, precision, recall, and F1-score. It is worth noting the accuracy of more than 0.96 obtained for the pixel classification of the affected and healthy leaves, and the good detection ability of the probabilistic classification model, which reached an accuracy of 0.87 for infected palm trees. The proposed methodology is presented as an efficient tool for identifying infected palm specimens, using spectral information, reducing the need for fieldwork and facilitating phytosanitary treatment.

Active Bayesian Deep Learning With Vector Sensor for Passive Sonar Sensing of the Ocean

Multiple studies have explored and reported on using passive sonar scalar pressure fields for both biological and man-made target classification. Here, inclusion of the vector field is considered and evaluated as an input feature in the classification of ship noises. We find that the inclusion of vector field information can significantly improve the understanding of the classification model performance. Based on an analysis of the epistemic and aleatoric uncertainties of probabilistic classification models, using vector field information as additional inputs leads to an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F_{1}$</tex-math></inline-formula> score of 0.8 in classification, comparable to scalar pressure field inputs only, but with 39% lower epistemic uncertainty in favor of utilizing the vector field component. Moreover, we verify that a similar conclusion applies to the active learning scenario, where we demonstrate the ability to utilize vector field information and epistemic uncertainty estimates to train a model that achieves an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F_{1}$</tex-math></inline-formula> score of 0.8 while using only 23% of the overall data. The ability to decompose uncertainty into an aleatoric and epistemic component leads to additional model explainability.

Discrimination of Copper Molten Marks through a Fire Reproduction Experiment Using Microstructure Features.

The copper molten marks at a fire site provide important clues for determining the causes of fire. Four factors have been presented to quantitatively discriminate copper molten marks, namely the fraction of (001) component perpendicular to the demarcation line, the grain aspect ratio, the fraction of Σ3 boundaries, and the fraction of maximum grain size. However, only laboratory-level results of these parameters have been presented, and their applicability in actual fires is yet to be verified. In this study, a fire reproduction experimental system was configured to generate molten marks similar to those in actual fire sites. The molten marks were measured by electron backscatter diffraction and applied to the four discriminant factors. The results obtained similar characteristics to those of the laboratory unit, confirming the applicability of the four discriminant factors. Discriminant equations and processes that can distinguish the primary and secondary arc beads were derived using the molten marks generated in the laboratory and reproduction experiments. Furthermore, a probabilistic discrimination method and classification model developed by machine learning were proposed. Therefore, the use of the discriminants in actual fires can improve the reliability of the statistics and prevent the recurrence of similar fires.

Predicting the probability distribution of bus travel time to measure the reliability of public transport services

An important aspect of the quality of a public transport service is its reliability, which is defined as the invariability of the service attributes. In order to measure reliability during the service planning phase, a key piece of information is the long-term prediction of the density of the travel time, which conveys the uncertainty of travel times. This work empirically compares probabilistic models for the prediction of the conditional probability density function (PDF) of the travel time and proposes a simulation framework taking as input the latter distributions to approximate the expected secondary delays, a measure of the reliability of public transport services. Two types of probabilistic models, namely similarity-based density estimation models and a smoothed logistic regression for probabilistic classification model, are compared on a dataset of more than 41,000 trips and 50 bus routes of the city of Montréal. A similarity-based density estimation model using a k-nearest neighbors method and a log-logistic distribution predicted the best estimate of the true conditional PDF of the travel time and generated the most accurate approximations of the expected secondary delays on this dataset. This model reduced the mean squared error of the expected secondary delay by approximately 9% compared to the benchmark model, namely a random forest. This result highlights the added value of modeling the conditional PDF of the travel time with probabilistic models.

A Novel Machine Learning Based Probabilistic Classification Model for Heart Disease Prediction

Cardiovascular disease (CVD) is most dreadful disease that results in fatal-threats like heart attacks. Accurate disease prediction is very essential and machine-learning techniques contribute a major part in predicting occurrence. In this paper, a novel machine learning based model for accurate prediction of cardiovascular disease is developed that applies unique feature selection technique called Chronic Fatigue Syndrome Best Known Method (CFSBKM). Each feature is ranked based on the feature importance scores. The new learning model eliminates the most irrelevant and low importance features from the datasets thereby resulting in the robust heart disease risk prediction model. The multi-nominal Naive Bayes classifier is used for the classification. The performance of the CFSBKM model is evaluated using the Benchmark dataset Cleveland dataset from UCI repository and the proposed models out-perform the existing techniques.

Multi-Sensor Data Driven with PARAFAC-IPSO-PNN for Identification of Mechanical Nonstationary Multi-Fault Mode

Data analysis has wide applications in eliminating the irrelevant and redundant components in signals to reveal the important informational characteristics that are required. Conventional methods for multi-dimensional data analysis via the decomposition of time and frequency information that ignore the information in signal space include independent component analysis (ICA) and principal component analysis (PCA). We propose the processing of a signal according to the continuous wavelet transform and the construction of a three-dimensional matrix containing the time–frequency–space information of the signal. The dimensions of the three-dimensional matrix are reduced by parallel factor analysis, and the time characteristic matrix, frequency characteristic matrix, and spatial characteristic matrix are obtained with tensor decomposition. Through the comparative analysis of the simulation and the experiment, the time characteristic matrix and the frequency characteristic matrix can accurately characterize the normal and fault states of the mechanical equipment. On this basis, the authors established a probabilistic neural network classification model optimized by the improved particle swarm algorithm (IPSO). The parallel factor (PARAFAC) decomposition algorithm can extract features from the centrifugal pump experimental data for normal and multiple fault states, establish the mapping relationship of different fault features of the centrifugal pump in time, frequency, and space, and import the fault features into the model classification. The above measures can significantly improve the fault identification rate and accuracy for a centrifugal pump.

A Deep Probabilistic Sensing and Learning Model for Brain Tumor Classification With Fusion-Net and HFCMIK Segmentation.

Goal: Implementation of an artificial intelli gence-based medical diagnosis tool for brain tumor classification, which is called the BTFSC-Net. Methods: Medical images are preprocessed using a hybrid probabilistic wiener filter (HPWF) The deep learning convolutional neural network (DLCNN) was utilized to fuse MRI and CT images with robust edge analysis (REA) properties, which are used to identify the slopes and edges of source images. Then, hybrid fuzzy c-means integrated k-means (HFCMIK) clustering is used to segment the disease affected region from the fused image. Further, hybrid features such as texture, colour, and low-level features are extracted from the fused image by using gray-level cooccurrence matrix (GLCM), redundant discrete wavelet transform (RDWT) descriptors. Finally, a deep learning based probabilistic neural network (DLPNN) is used to classify malignant and benign tumors. The BTFSC-Net attained 99.21% of segmentation accuracy and 99.46% of classification accuracy. Conclusions: The simulations showed that BTFSC-Net outperformed as compared to existing methods.

Controllable and non-controllable factors to measure performance in primary care practices under Medicare alternative payment models

We analyze the two recent Medicare alternative payment models, the comprehensive primary care plus (CPC+) and the primary care first (PCF). Both models comprise fee-for-service, traditional capitation, and pay-for-performance (P4P) components. The main objective of these reimbursement models is to advance toward value-based care. However, the models confer some hesitations since the P4P component is based on factors not entirely controlled by the practice, increasing the potential admission of healthier patients and affecting the profit of small primary care practices. We have modified the P4P component in both models to include a non-controllable agent (the hierarchical condition category score) and a controllable factor (the Bice–Boxerman continuity of care index) through a probabilistic classification model to predict hospital admissions. This study aims to determine the impact of adjusting the P4P component, in the CPC+ and PCF reimbursement models, on the profit per team, revenue for performance per team, and severity of admitted patients. We develop a mixed-integer programming formulation and analyze, using a 2k factorial design, the reimbursement models and the main elements of their adjusted P4P components (i.e., the probabilistic classification model coefficients and hospital admission threshold). The results indicate that the coefficients of the probabilistic classification model and the hospital admission threshold have a significant effect on the profit and revenue for performance per team. There is also a tendency of the PCF to admit less severe patients than the CPC+. Yet, the effects are more notable in the PCF payment model because the proportion of P4P in the total revenue under the CPC+ is minimal (16.5% versus <1%). Similarly, the PCF’s downside is its sensitivity to P4P changes, displaying high variability in the output variables under analysis.

Controls of outbursts of moraine-dammed lakes in the greater Himalayan region

Abstract. Glacial lakes in the Hindu Kush–Karakoram–Himalayas–Nyainqentanglha (HKKHN) region have grown rapidly in number and area in past decades, and some dozens have drained in catastrophic glacial lake outburst floods (GLOFs). Estimating regional susceptibility of glacial lakes has largely relied on qualitative assessments by experts, thus motivating a more systematic and quantitative appraisal. Before the backdrop of current climate-change projections and the potential of elevation-dependent warming, an objective and regionally consistent assessment is urgently needed. We use an inventory of 3390 moraine-dammed lakes and their documented outburst history in the past four decades to test whether elevation, lake area and its rate of change, glacier-mass balance, and monsoonality are useful inputs to a probabilistic classification model. We implement these candidate predictors in four Bayesian multi-level logistic regression models to estimate the posterior susceptibility to GLOFs. We find that mostly larger lakes have been more prone to GLOFs in the past four decades regardless of the elevation band in which they occurred. We also find that including the regional average glacier-mass balance improves the model classification. In contrast, changes in lake area and monsoonality play ambiguous roles. Our study provides first quantitative evidence that GLOF susceptibility in the HKKHN scales with lake area, though less so with its dynamics. Our probabilistic prognoses offer improvement compared to a random classification based on average GLOF frequency. Yet they also reveal some major uncertainties that have remained largely unquantified previously and that challenge the applicability of single models. Ensembles of multiple models could be a viable alternative for more accurately classifying the susceptibility of moraine-dammed lakes to GLOFs.

A probabilistic clustering model for hate speech classification in twitter

The key challenges for automatic hate-speech classification in Twitter are the lack of generic architecture, imprecision, threshold settings and fragmentation issues. Most studies used binary classifiers for hate speech classification, but these classifiers cannot really capture other emotions that may overlap between positive or negative class. Hence, a probabilistic clustering model for hate speech classification in twitter was developed to tackle problems with hate speech classification. A metadata extractor was used to collect tweets containing hate speech keywords and a crowd-sourced experts was employed to label the collected hate tweets into two categories: hate speech and non-hate speech. Features representation was done with Term Frequency- Inverse Document Frequency (TF-IDF) model and enhanced with topics inferred by a Bayes classifier. A rule-based clustering method was used to automatically classify real-time tweets into the correct topic clusters. Fuzzy logic was then used for hate speech classification using semantic fuzzy rules and a score computation module. From the evaluation results, it was observed that the developed model performed better in hate speech detection with F1-sore of 0.9256 using a 5-fold cross validation. Similarly, the developed model for hate speech classification performed better with F1-score of 91.5 compared to related models. The developed model also indicates a more perfect test having an AUC of 0.9645, when compared to similar methods. The Paired Sample t-Test validated the efficiency of the developed model for hate speech classification.

Probabilistic classification models for the in situ authentication of iberian pig carcasses using near infrared spectroscopy

Iberian pig ham is one of several high value European food products that are the subject of significant attempts at fraud because of the high price differences between commercial categories. Iberian pig products are classified by the Spanish regulations into different categories, mainly depending on the feeding regime during the fattening phase and the race involved, being of Premium quality those products obtained from the animals fed with acorns and other natural resources. Most of the previous NIRS studies related to the Iberian pig have involved the use of at-line instruments to predict quantitative quality parameters. This paper explores the use of the NIR spectra (369 for training and 199 for validation) to classify samples according to the categories Premium (animals fed with acorn) and Non Premium (animals fed with compound feeds), using a MicroNIR™ Pro1700 microspectrometer to analyse individual carcasses in situ at the slaughterhouse line. Four discriminant methods were explored: linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), Kernel Bayes and Logistic Regression. These are all discriminant methods that naturally produce classification probabilities to quantify the uncertainty of the results. Rules were tuned and methods compared using both classification error rates and a probability scoring rule. LDA gave the best results, attaining an overall accuracy of 93% and providing well-calibrated classification probabilities.

Probabilistic soil classification and stratification in a vertical cross-section from limited cone penetration tests using random field and Monte Carlo simulation

Classification and stratification (or zonation) of subsurface soils are important tasks in geotechnical site investigation. Due to the limit of time, budget, or access to subsurface soils, subsurface soil information obtained from investigation points (e.g., boreholes, cone penetration tests (CPTs)) in a specific site is often limited (e.g., a few boreholes or CPT soundings), resulting in great challenge in interpretation of the site investigation data obtained and significant uncertainty in the inferred subsurface soil classification and stratification. A novel probabilistic method is developed in this paper for properly accounting for the uncertainty associated with CPT-based subsurface soil classification and stratification. The method properly classifies and stratifies subsurface soils in a 2D vertical cross-section from limited CPT soundings. A limited number of 1D CPT sounding data is firstly interpolated to produce a 2D vertical cross-section, and the associated interpolation uncertainty is modelled explicitly using random field theory. Probabilistic soil classification model is also developed to account for the uncertainty associated with the empirical soil behavior type classification model. Then, the interpolation uncertainty and soil classification model uncertainty are considered simultaneously in a Monte Carlo simulation framework. Both simulated and real data examples are used to illustrate the proposed method. The results indicate that the proposed method well predicts subsurface soil classification and stratification in a 2D vertical cross-section from limited CPT soundings, and properly quantifies the associated uncertainty. In addition, sensitivity studies on interpolation uncertainty and soil classification model uncertainty are performed.

Local probabilistic model for Bayesian classification: A generalized local classification model

In Bayesian classification, it is important to establish a probability distribution model, e.g., a Gaussian distribution for each class for probability estimation. Most of the previous methods modeled the probability distribution in the whole sample space. However, real-world problems are usually too complex to model in the whole sample space; some fundamental assumptions are required to simplify the global model, for example, the class conditional independence assumption for naive Bayesian classification. In this paper, with the insight that the distribution in a local sample space should be simpler than that in the whole sample space, a local probabilistic model established for a local region is expected much simpler and can relax the fundamental assumptions that may not be true in the whole sample space. Based on these advantages we propose establishing local probabilistic models for probability estimation in Bayesian classification. In addition, a Bayesian classifier adopting a local probabilistic model can even be viewed as a generalized local classification model; by tuning the size of the local region and the corresponding local model assumption, a fitting model can be established for a particular classification problem. The experimental results on several real-world datasets demonstrate the effectiveness of local probabilistic models for Bayesian classification.

Profiling identification with Europe and the EU project in the European regions

Recent political events in the European Union (EU) highlighted a growing dissatisfaction of citizens in several EU regions with the EU institutions’ management of socio-economic and financial challenges. This eventually led to a political legitimization crisis, whose drivers are partially shared among EU regions and partially area-specific. However, the relation between citizens’ identification with the EU project and the regions’ characteristics has not been analysed yet. In this article, we fill in this gap by addressing tree research questions: i) To what extent do EU citizens identify with Europe and the EU project? ii) Do European regions have different patterns and level of identification?, and iii) Are the results driven by specific socio-economic variables?&#x0D; Answering these questions is crucial to inform a more inclusive and resilient design of the EU Cohesion Policy in a crucial period for reforming the EU. To this purpose, we develop a novel probabilistic classification model, IdentEU, which embeds with the concept of individual identification with Europe. We use micro-level data from a survey implemented within the PERCEIVE project. We find that the influencing variables that mostly affect (citizens and) regions’ identification with the European project are trust in the EU institutions; the effectiveness of EU Cohesion Policy and spending; the level of corruption. These issues gain relevance at the light of three main challenges that affected the EU socio-economic development path in the last decade, i.e. the 2008 financial crisis, the globalization process, and Brexit.

LCBM: A Multi-View Probabilistic Model for Multi-Label Classification.

Multi-label classification is an important research topic in machine learning, for which exploiting label dependencies is an effective modeling principle. Recently, probabilistic models have shown great potential in discovering dependencies among labels. In this paper, motivated by the recent success of multi-view learning to improve the generalization performance, we propose a novel multi-view probabilistic model named latent conditional Bernoulli mixture (LCBM) for multi-label classification. LCBM is a generative model taking features from different views as inputs, and conditional on the latent subspace shared by the views a Bernoulli mixture model is adopted to build label dependencies. Inside each component of the mixture, the labels have a weak correlation which facilitates computational convenience. The mean field variational inference framework is used to carry out approximate posterior inference in the probabilistic model, where we propose a Gaussian mixture variational autoencoder (GMVAE) for effective posterior approximation. We further develop a scalable stochastic training algorithm for efficiently optimizing the model parameters and variational parameters, and derive an efficient prediction procedure based on greedy search. Experimental results on multiple benchmark datasets show that our approach outperforms other state-of-the-art methods under various metrics.

Location based Web Object Search using Probabilistic Classification Model.

The classical Web search engines focus on satisfying the information need of the users by retrieving relevant Web documents corresponding to the user query. The Web document contains the information on different Web objects such as authors, automobiles, political parties e.t.c. The user might be accessing the Web document to procure information about a specific Web object, the remaining information in the Web object [2-6] becomes redundant specific to the user. If the size of Web documents is significantly large and the user information requirement is small fraction of the document, the user has to invest effort in locating the required information inside the document. It would be much more convenient if the user is provided with only the required Web object information located inside the Web documents. Web object search engines provide Web search facility through vertical search on Web objects. In this paper the main goal we considered is the objective information present in different documents is extracted and integrated into an object repository over which the Web object search facility is built.