Parsimonious Minimal Learning Machine via Multiresponse Sparse Regression.

Landslide susceptibility analysis is beneficial information for a wide range of applications, including land use management plans. The present attempt has shed light on an efficient landslide susceptibility mapping framework that involves an adaptive neural-fuzzy inference system (ANFIS), which incorporates three metaheuristic methods including grey wolf optimization (GWO), particle swarm optimization (PSO), and shuffled frog leaping algorithm (SFLA) in the East Azerbaijan of Iran. To achieve this goal, 10 landslide occurrence-related influencing factors were pondered. A sum of 766 locations with landslide potentiality was recognized in the context of the study, and the Pearson correlation technique utilized in order to select the influencing factors in landslide models. The association between landslides and conditioning factors was also evaluated using a probability certainty factor (PCF) model. In the next phase, three data mining techniques were united with the ANFIS model, comprising ANFIS-grey wolf algorithm (ANFIS-GW), ANFIS-particle swarm optimization (ANFIS-PSO), and ANFIS-shuffled frog leaping algorithm (ANFIS-SFLA), were structured by the training dataset. Lastly, the receiver operating characteristic (ROC) and statistical procedures were utilized with the aim of validating and contrasting the predictive capability of the models. The findings of the study in terms of the Pearson correlation technique method for the importance ranking of conditioning factors in the context area uncovered that slope, aspect, normalized difference vegetation index (NDVI), and elevation have the highest impact on the occurrence of the landslide. All in all, the ANFIS-PSO model had high performance on both the training (RMSE = 0.288, MAE = 0.069, AUC = 0.89) and validation dataset (RMSE = 0.309, MAE = 0.097, AUC = 0.89), after which, the ANFIS-GWO model and the ANFIS-SFLA model demonstrated the second and third rates. Besides, the study revealed that benefiting the optimization algorithm with the proper selection of the techniques could facilitate landslide susceptibility modeling.

Right choice of meta-models is one of the most important factors determining the accuracy of predicting seawater intrusion phenomena in the aquifers of coastal belts. In this paper, entropy weight based decision theory is applied to rank the performances of meta-models. Six meta-models trained and validated by a set of input-output training patterns generated from a unified flow and solute transport model for saltwater intrusion are considered. Entropy weights are assigned to performance evaluation indicators in order to decide on the comparative significance of the indicators in meta-model performance. Meta-models are then ranked by incorporating this relative importance of individual performance indicators. This method of ranking provides reliability in meta-model selection by considering a set of performance indicators instead of relying on a single indicator. Furthermore, this method is compared with variation coefficient weighting method. It is shown that the proposed entropy weight based ranking methodology can be successfully applied to select the best meta-model for predicting seawater intrusion processes in coastline aquifers.

The emergence of 6G wireless networks presents new challenges, for which cell-free massive MIMO combined with machine learning (ML) offers a promising solution. A key requirement for practical deployment is the generalizability of ML models—their ability to maintain robust performance across varying propagation conditions, user distributions, and network topologies. However, achieving generalizability typically demands large, diverse training datasets and high model complexity, which can hinder practical feasibility. This study analyzes the robustness of a low-complexity deep neural network (DNN) trained for power control under a single network configuration. The model’s robustness is assessed by testing it across a wide range of unseen scenarios, including changes in the number of access points, user equipment, and propagation environments. The DNN is trained to emulate three power control schemes: max-min spectral efficiency (SE) fairness, sum SE maximization, and fractional power control. To rigorously evaluate robustness, we compare the cumulative distribution functions of performance metrics quantitatively using the Kolmogorov–Smirnov test. Results show strong robustness, particularly for the sum SE scheme, with D statistics below 0.05 and p-values above 0.001. This work provides a reproducible framework and dataset to support further research into practical ML-based power control in cell-free massive MIMO systems.

Probabilistic identification of surface recession patterns in heritage buildings based on digital photogrammetry

This paper proposes a query suggestion method combining two ranked retrieval methods: TF-IDF and Jaccard coefficient. Four performance criteria plus user evaluation have been adopted to evaluate this combined method in terms of ranking and relevance from different perspectives. Two experiments have been conducted using carefully designed eighty test queries which are related to eight topics. One experiment aims to evaluate the quality of the query suggestions generated by the proposed method, and the other aims to evaluate the improvement of the relevance of retuned documents in interactive web search by using thequery suggestions so as to evaluate the effectiveness of the developed method. The experimental results show that the method developed in this paper is the best method for query suggestion among the methods evaluated, significantly outperforming the most popularly used TF-IDF method. In addition, the query suggestions generated by the proposed method significantly improve the relevance of returned documents in interactive web search in terms of increasing the precision or the number of highly relevant documents.

Purposes and approach: Acute castration and/or tailing distress in lambs has been examined extensively during the last decade. At least 59 different approaches to assessing and alleviating this distress have been reported so that the literature is quite complex. The purpose of this paper is to provide an overview of the literature on castration and/or tailing distress, where the distress was assessed using acute changes in plasma cortisol concentrations. A method of analysis involving the integrated cortisol response (i.e. the areas under the cortisol curves while the plasma concentration is above pretreatment values) to each treatment and using treatments which were common to different studies as reference points, allowed meaningful comparison within and between studies. A 6-point ranking scale emerged, where rank 1 represented the least distress and rank 6 the most distress. Comparison of acute distress responses: This analysis revealed the following major points. Surgical methods of castration and/or tailing cause the greatest cortisol responses (rank 5 or 6). Most ring and ring plus clamp methods of castration plus tailing or castration, used without a local anaesthetic or systemic analgesic, cause rank 4 responses. One form of ring plus clamp castration (i.e. applying the clamp for 10 s across the full width of the scrotum distal to the ring in lambs aged no more than 1 week) reduces the cortisol response to rank 1.When these lambs are also tailed by applying a ring and clamp in a similar manner to the tail, they also exhibit a rank 1 response. Local anaesthetic given 10–20 or 1–2 min or 10–15 s before or immediately after ring only castration and/or tailing can virtually abolish the cortisol response (rank 1), depending on the site(s) of injection. For ring or ring plus clamp castration, the most effective sites (as judged by cortisol responses) are the neck of the scrotum or the testes. Delivery of local anaesthetic to achieve successful nerve blockade can be by needle, high-pressure needleless administration or, for the tail only, by an aerosol spray. Local anaesthetic injected into the scrotal neck, spermatic cords and/or testes has little effect on the overall cortisol response to clamp castration. Reductions in cortisol responses to clamp castration or to ring tailing can occur after administration of non-steroidal anti-inflammatory drugs. Most tailing methods elicit cortisol responses that are several ranks lower than those caused by castration plus tailing or castration alone. Although tailing by most methods elicits rank 1 cortisol responses, the use of local anaesthetic or non-steroidal anti-inflammatory drugs can reduce the response within the rank 1 range. Recommendation: Farmers should be encouraged to choose the lowest ranked method that is economically and practically feasible for them. Specific methods such as surgical castration should be discouraged.

Document images are now widely captured by handheld devices such as mobile phones. The OCR performance on these images are largely affected due to geometric distortion of the document paper, diverse camera positions and complex backgrounds. In this paper, we propose a simple yet effective approach to rectify distorted document image by estimating control points and reference points. After that, we use interpolation method between control points and reference points to convert sparse mappings to backward mapping, and remap the original distorted document image to the rectified image. Furthermore, control points are controllable to facilitate interaction or subsequent adjustment. We can flexibly select post-processing methods and the number of vertices according to different application scenarios. Experiments show that our approach can rectify document images with various distortion types, and yield state-of-the-art performance on real-world dataset. This paper also provides a training dataset based on control points for document dewarping. Both the code and the dataset are released at https://github.com/gwxie/Document-Dewarping-with-Control-Points.

Normalization is a prerequisite for almost all follow-up steps in microarray data analysis. Accurate normalization across different experiments and phenotypes assures a common base for comparative yet quantitative studies using gene expression data. In this paper, we report a comparison study of four normalization approaches, namely, linear regression (LR), Loess regression, invariant ranking (IR) and iterative nonlinear regression (INR), for gene expression data normalization. Among these four methods, LR and Loess regression methods use all available genes to estimate either a linear or nonlinear normalization function, while IR and INR methods feature some iterative processes to identify invariantly expressed genes (IEGs) for nonlinear normalization. We tested these normalization approaches on three real microarray data sets and evaluated their performances in terms of variance reduction and fold-change preservation. By comparison, we found that (1) LR method exhibited the worst performance in both variance reduction and fold-change preservation, and (2) INR method showed an improved performance in achieving low expression variance across replicates and excellent fold-change preservation for differentially expressed genes.

Multimodal biometrics is often used as it solves the issues of unimodal biometrics like non-universality, noisy data and security. Face and palmprint of a person do not change easily with time. To extract both local and global features of an image, we propose a partition based DWT (Discrete Wavelet Transform) and 2DPCA (2Dimensional Principal Component Analysis) for feature extraction. The fusion of face and palm print is done at rank level using Highest rank, Borda count and Logistic Regression methods. The proposed method is better than 2DPCA (upto 1% higher rate) and DWT (upto 3% higher rate) in terms of average recognition rate for Unimodal systems. For Multimodal approach, the proposed method gives an equal error rate of 0.08 when compared to Unimodal approach of 0.24 for palm and 0.16 for face.

Received Signal Strength (RSS) fingerprinting based WiFi localization systems has been widely studied in recent years for indoor scenarios, whose accuracy is closely related to the reliability of the database. To attain satisfactory performance, the database needs to be updated timely to compensate for changes in the environment. In this work, we propose a rank fingerprinting based localization system, in which the radio map only takes account of the most consistent access points (APs) with most meaningful information, and uses the rank method instead of the signal strength with clustering the most correlated reference point to estimate the position. In order to keep a reasonable size of the radio map, a dimension control strategy has been proposed. The proposed algorithm deals with manifold devices to achieve high room-level accuracy in a shopping mall. It has been compared with two present algorithms, the rank based fingerprinting (RBF) algorithm and the support vector machine (SVM) based algorithm, to show its better performance.

The implementation of mobile users' emotional perception is complex and inefficient at present. Therefore, we propose a method that realizes emotion analysis by exploring the relationship between user emotion and time characteristics, using natural language processing, SVM, and mathematical analysis methods. Then combined the corpus built by ourselves, and the CHI feature extraction method is used to extract the feature from the training data set, the data is transformed into the feature matrix according to the feature values, and the training set is modeled and trained by the SVM method, and then optimize the parameters using genetic algorithm. According to the result of the SVM decision function, the weighted idea is used to weight the data, and its weight is modeled and calculated according to the weight loss mechanism we proposed, and the final result will be obtained as a condition for the determination of emotion perception.

Purpose Defining parameters for typical development in bilingual children's first and second languages can serve as the basis for accurate language assessment. This is the first study to characterize Vietnamese and English grammatical development in a sample of bilingual children. Method Participants were 89 Vietnamese-English bilingual children, aged 3-8 years. Children completed story retell tasks in Vietnamese and English. Stories were transcribed and analyzed for grammaticality, error patterns, subordination index, and types of subordinating clauses. Of key interest were associations with age and identifying developmental patterns that were shared across languages or unique to a given language. Results Age correlated with more measures in English than in Vietnamese, suggesting that older children had higher grammaticality and greater syntactic complexity in English than younger children. Children also produced greater syntactic complexity with age in Vietnamese, but not higher grammaticality. There were a set of error patterns shared across languages (e.g., object omission) and patterns specific to each language (e.g., classifier errors in Vietnamese, tense errors in English). While children produced nominal, adverbial, and relative clauses in Vietnamese and English, the proportion of each clause type differed by language. Conclusions Results from this typically developing sample provide a reference point to improve clinical practice. Characterizing developmental patterns in sentence structure in Vietnamese and English lays the groundwork for investigations of language disorders in this bilingual population.

A person’s gait is a behavioral trait that is uniquely associated with each individual and can be used to recognize the person. As information about the human gait can be captured by wearable devices, a few studies have led to the proposal of methods to process gait information for identification purposes. Despite recent advances in gait recognition, an open set gait recognition problem presents challenges to current approaches. To address the open set gait recognition problem, a system should be able to deal with unseen subjects who have not included in the training dataset. In this paper, we propose a system that learns a mapping from a multimodal time series collected using insole to a latent (embedding vector) space to address the open set gait recognition problem. The distance between two embedding vectors in the latent space corresponds to the similarity between two multimodal time series. Using the characteristics of the human gait pattern, multimodal time series are sliced into unit steps. The system maps unit steps to embedding vectors using an ensemble consisting of a convolutional neural network and a recurrent neural network. To recognize each individual, the system learns a decision function using a one-class support vector machine from a few embedding vectors of the person in the latent space, then the system determines whether an unknown unit step is recognized as belonging to a known individual. Our experiments demonstrate that the proposed framework recognizes individuals with high accuracy regardless they have been registered or not. If we could have an environment in which all people would be wearing the insole, the framework would be used for user verification widely.

This paper focuses on the problem of risk assessment method for large scale sports events which is an important problem in modern sports management. The index system for large scale sports events risk assessment is proposed in advance, which is made up of three categories: 1) risk before match, 2) risk in match and 3) risk after match. Particularly, eighteen influencing factors are design which can cover all aspects of the large scale sports events risk assessment. Structure of the hybrid neural network is contructed by three layers, which are "the input layer", "the hidden layer", and "the output layer". Particularly, the output layers can compute the regression function values, and the regression function of the artificial neural network can be computed through a linear integration of some nonlinear basis functions. Afterwards, utilizing a training dataset and its updating version, values of the decision functions of the multi-class support vector classifier can be obtained by the regression functions based on the artificial neural network. Finally, experiments are conducted to test the effectiveness of our algorithm. The conclusions can be drawn that compared with the artificial neural network, the proposed hybrid neural network is more suitable for the risk assessment of large scale sports events.

Decision trees are discriminative classifiers that hierarchically partition the input space to achieve regions containing instances having uniform class label. Existing works in this area have mostly focused on C4.S trees that learn axis aligned partitions. On the other hand, neural trees learn oblique partitions from data and use lesser number of decision nodes hosting perceptrons. However, these perceptrons are susceptible to data imbalances. This motivated us to propose a progressively balanced neural tree where training dataset are balanced prior to perceptron learning. The second contribution is the optimization of the decision function with respect to entropy impurity based objective functions. This formulation also allows a parent node to have more than two child nodes. The proposed algorithm is benchmarked on ten standard datasets against three baseline multi-class classification algorithms.