Traditional Statistical Classification Research Articles

High-precision bladder cancer diagnosis method: 2D Raman spectrum figures based on maintenance technology combined with automatic weighted feature fusion network

BackgroundRaman spectroscopy has been extensively utilized as a marker-free detection method in the complementary diagnosis of cancer. Multivariate statistical classification analysis is frequently employed for Raman spectral data classification. Nevertheless, traditional multivariate statistical classification analysis performs poorly when analyzing large samples and multicategory spectral data. In addition, with the advancement of computer vision, convolutional neural networks (CNNs) have demonstrated extraordinarily precise analysis of two-dimensional image processing. ResultCombining 2D Raman spectrograms with automatic weighted feature fusion network (AWFFN) for bladder cancer detection is presented in this paper. Initially, the s-transform (ST) is implemented for the first time to convert 1D Raman data into 2D spectrograms, achieving 99.2% detection accuracy. Second, four upscaling techniques, including short time fourier transform (STFT), recurrence map (RP), markov transform field (MTF), and grammy angle field (GAF), were used to transform the 1D Raman spectral data into a variety of 2D Raman spectrograms. In addition, a particle swarm optimization (PSO) algorithm is combined with VGG19, ResNet50, and ResNet101 to construct a weighted feature fusion network, and this parallel network is employed for evaluating multiple spectrograms. Class activation mapping (CAM) is additionally employed to illustrate and evaluate the process of feature extraction via the three parallel network branches. The results demonstrate that the combination of a 2D Raman spectrogram along with a CNN for the diagnosis of bladder cancer obtains a 99.2% accuracy rate，which indicates that it is an extremely promising auxiliary technology for cancer diagnosis. SignificanceThe proposed two-dimensional Raman spectroscopy method has an improved precision than one-dimensional spectroscopic data, which presents a potential methodology for assisted cancer detection and providing crucial technical support for assisted diagnosis.

A voice search engine for military symbols to enhance the drafting of operational plan documents on digital map

The process of searching for information to serve the construction of operational plan documents on a digital map is still being done manually and needs to be automated in order to improve efficiency. Speech recognition and natural language processing technologies, commonly used in chatbots, virtual assistants, voice commands, and voice search, could be promising tools to overcome this problem. This paper proposes a framework for deploying a voice search engine that uses Whisper, a deep learning-based automatic speech recognition model, and combines TF-IDF, N-gram, and Truncated SVD as feature extraction approaches to search for text ground truth in a dictionary of military symbols using Cosine similarity. Despite the small size of a custom dataset, the experiments show promising results, achieving an accuracy of 82.00%. Our achievement surpasses that of several traditional statistical methods and classification models.

A machine learning approach for hierarchical classification of software requirements

Context:Classification of software requirements into different categories is a critically important task in requirements engineering (RE). Developing machine learning (ML) approaches for requirements classification has attracted great interest in the RE community since the 2000s. Objective:This paper aims to address two related problems that have been challenging real-world applications of ML approaches: the problems of class imbalance and high dimensionality with low sample size data (HDLSS). These problems can greatly degrade the classification performance of ML methods. Methods:The paper proposes HC4RC, a novel ML approach for multiclass classification of requirements. HC4RC solves the aforementioned problems through semantic-role based feature selection, dataset decomposition and hierarchical classification. We experimentally compare the effectiveness of HC4RC with three closely related approaches — two of which are based on a traditional statistical classification model whereas one using an advanced deep learning model. Results:Our experiment shows: (1) The class imbalance and HDLSS problems present a challenge to both traditional and advanced ML approaches. (2) The HC4RC approach is simple to use and can effectively address the class imbalance and HDLSS problems compared to similar approaches. Conclusion:This paper makes an important practical contribution to addressing the class imbalance and HDLSS problems in multiclass classification of software requirements.

Music Morphology Interaction under Artificial Intelligence in Wireless Network Environment.

Music has become the main information carrier, and music and its emotional expression are accurately classified to obtain relevant information. However, how to classify music accurately is a problem that needs to be discussed. The concept and feature extraction strategy of the morphology of music are described. Moreover, the feature extraction and morphological classification elements of digital music are introduced. Next, music morphology is recognized and classified based on the neural network and relief algorithm. In the network, by randomly selecting different music types, the audio data is input into the neural network as the original data and processed by the relief algorithm. The classification and recognition accuracy of the Relief algorithm are verified by changing the number of iterations. The results show that the model's classification accuracy based on the number of iterations is 78.958%. Then, the traditional statistical analysis classification method's performance is compared with the proposed model. The recognition accuracy of the model proposed reaches 92%, which shows that the model can effectively classify music morphology. This study provides a theoretical basis for music morphology recognition in the wireless network environment.

Sentiment Analysis of Sinhala News Comments

Sinhala is a low-resource language, for which basic language and linguistic tools have not been properly defined. This affects the development of NLP-based end-user applications for Sinhala. Thus, when implementing NLP tools such as sentiment analyzers, we have to rely only on language-independent techniques. This article presents the use of such language-independent techniques in implementing a sentiment analysis system for Sinhala news comments. We demonstrate that for low-resource languages such as Sinhala, the use of recently introduced word embedding models as semantic features can compensate for the lack of well-developed language-specific linguistic or language resources, and text classification with acceptable accuracy is indeed possible using both traditional statistical classifiers and Deep Learning models. The developed classification models, a corpus of 8.9 million tokens extracted from Sinhala news articles and user comments, and Sinhala Word2Vec and fastText word embedding models are now available for public use; 9,048 news comments annotated with POSITIVE/NEGATIVE/NEUTRAL polarities have also been released.

Bioimage-Based Prediction of Protein Subcellular Location in Human Tissue with Ensemble Features and Deep Networks.

Prediction of protein subcellular location has currently become a hot topic because it has been proven to be useful for understanding both the disease mechanisms and novel drug design. With the rapid development of automated microscopic imaging technology in recent years, classification methods of bioimage-based protein subcellular location have attracted considerable attention for images can describe the protein distribution intuitively and in detail. In the current study, a prediction method of protein subcellular location was proposed based on multi-view image features that are extracted from three different views, including the four texture features of the original image, the global and local features of the protein extracted from the protein channel images after color segmentation, and the global features of DNA extracted from the DNA channel image. Finally, the extracted features were combined together to improve the performance of subcellular localization prediction. From the performance comparison of different combination features under the same classifier, the best ensemble features could be obtained. In this work, a classifier based on Stacked Auto-encoders and the random forest was also put forward. To improve the prediction results, the deep network was combined with the traditional statistical classification methods. Stringent cross-validation and independent validation tests on the benchmark dataset demonstrated the efficacy of the proposed method.

New soft biometrics for limited resource in keystroke dynamics authentication

In recent years, some researches on free text authentication by keystroke dynamics have been proposed. The main problem of these proposed researches is the requirement of a very long training time. To increase users’ willingness to use the proposed authentication system, one simple and feasible way is shortening the training time. In this case, the training data is limited and is known as the limited resource problem. In this paper, we propose new soft biometrics and a new classifier for limited resources in free text authentication in English. Our new soft biometrics combines the idea of data mining and statistical prediction. Because our soft biometric is a mining result of limited resources, it is sensitive to outliers and a traditional statistical classifier cannot be applied. To solve this problem, the proposed classifiers considered the problem of an outlier and calculated the difference of cluster distribution. There are 114 participants in our experiments. Experimental results show that our approach can improve the accuracy of free text authentication in the case of limited resources.

Attention augmentation with multi-residual in bidirectional LSTM

Recurrent neural networks (RNNs) have been proven to be efficient in processing sequential data. However, the traditional RNNs have suffered from the gradient diminishing problem until the advent of Long Short-Term Memory (LSTM). However, LSTM is weak in capturing long-time dependency in sequential data due to the inadequacy of memory capacity in LSTM cells. To address this challenge, we propose an Attention-augmentation Bidirectional Multi-residual Recurrent Neural Network (ABMRNN) to overcome the deficiency. We propose an algorithm which integrates both past and future information at every time step with omniscient attention model. The multi-residual mechanism has also been leveraged in the proposed model targeting the pattern of the relationship between current time step and further distant time steps instead of only one previous time step. The results of experiments show that our model outperforms the traditional statistical classifiers and other existing RNN architectures.

Modeling dynamics of peri-urban interface based on principal component analysis (PCA) and cluster analysis (CA): a study of English Bazar Urban Agglomeration, West Bengal

Outward growth of large and medium size cities brings gradual changes in landuse/landcover which transforms the surrounding rural area to semi-urban or peri-urban area. But there is no such consensus regarding the definition of peri-urban region and its causes. It is normally defined as the settlements beyond, about or around core cities. These areas are characterized by the spillover developments of the core cities. Therefore, the present study attempts to develop suitable method to explore the growth dynamics of changing peri-urban interface with the help of some spatial and non-spatial data of peri-urban areas of English Bazar Urban Agglomeration (U.A). The study employs different techniques such as traditional statistical techniques (multivariate analysis), PCA classification and spatial analysis followed by cluster analysis. Thirteen buffer zones were created around the city with each of 1 km width. The final result shows that, the first six buffer zones are differing from the other zones in terms of different characteristics. Therefore, these six kilometer buffer zones are considered as peri-urban areas of English Bazar city.

Mapping mangrove forests using multi-tidal remotely-sensed data and a decision-tree-based procedure

Mangrove forests grow in intertidal zones in tropical and subtropical regions and have suffered a dramatic decline globally over the past few decades. Remote sensing data, collected at various spatial resolutions, provide an effective way to map the spatial distribution of mangrove forests over time. However, the spectral signatures of mangrove forests are significantly affected by tide levels. Therefore, mangrove forests may not be accurately mapped with remote sensing data collected during a single-tidal event, especially if not acquired at low tide. This research reports how a decision-tree −based procedure was developed to map mangrove forests using multi-tidal Landsat 5 Thematic Mapper (TM) data and a Digital Elevation Model (DEM). Three indices, including the Normalized Difference Moisture Index (NDMI), the Normalized Difference Vegetation Index (NDVI) and NDVIL·NDMIH (the multiplication of NDVIL by NDMIH, L: low tide level, H: high tide level) were used in this algorithm to differentiate mangrove forests from other land-cover and land-use types in Fangchenggang City, China. Additionally, the recent Landsat 8 OLI (Operational Land Imager) data were selected to validate the results and compare if the methodology is reliable. The results demonstrate that short-term multi-tidal remotely-sensed data better represent the unique nearshore coastal wetland habitats of mangrove forests than single-tidal data. Furthermore, multi-tidal remotely-sensed data has led to improved accuracies using two classification approaches: i.e. decision trees and the maximum likelihood classification (MLC). Since mangrove forests are typically found at low elevations, the inclusion of elevation data in the two classification procedures was tested. Given the decision-tree method does not assume strict data distribution parameters, it was able to optimize the application of multi-tidal and elevation data, resulting in higher classification accuracies of mangrove forests. When using multi-source data of differing types and distributions to map mangrove forests, a decision-tree method appears to be superior to traditional statistical classifiers.

Vascular segmentation of head phase-contrast magnetic resonance angiograms using grayscale and shape features

Background and objectiveIn neurosurgery planning, vascular structures must be predetermined, which can guarantee the security of the operation carried out in the case of avoiding blood vessels. In this paper, an automatic algorithm of vascular segmentation, which combined the grayscale and shape features of the blood vessels, is proposed to extract 3D vascular structures from head phase-contrast magnetic resonance angiography dataset. MethodsFirst, a cost function of mis-segmentation is introduced on the basis of traditional Bayesian statistical classification, and the blood vessel of weak grayscale that tended to be misclassified into background will be preserved. Second, enhanced vesselness image is obtained according to the shape-based multiscale vascular enhancement filter. Third, a new reconstructed vascular image is established according to the fusion of vascular grayscale and shape features using Dempster–Shafer evidence theory; subsequently, the corresponding segmentation structures are obtained. Finally, according to the noise distribution characteristic of the data, segmentation ratio coefficient, which increased linearly from top to bottom, is proposed to control the segmentation result, thereby preventing over-segmentation. ResultsExperiment results show that, through the proposed method, vascular structures can be detected not only when both grayscale and shape features are strong, but also when either of them is strong. Compared with traditional grayscale feature- and shape feature-based methods, it is better in the evaluation of testing in segmentation accuracy, and over-segmentation and under-segmentation ratios. ConclusionsThe proposed grayscale and shape features combined vascular segmentation is not only effective but also accurate. It may be used for diagnosis of vascular diseases and planning of neurosurgery.

PENGGUNAAN MOMEN INVARIANT, ECCENTRICITY, DAN COMPACTNESS UNTUK KLASIFIKASI MOTIF BATIK DENGAN K-NEAREST NEIGHBOUR

Batik classification which have diverse motifs need to be done to distinguish a pattern with another. In this paper, we present batik motifs(Ceplok, Parang, Semen, and Nitik) classification using Hu Moment Invariants, Eccentricity, and Compactness feature description. In classification stage, K-nearest neighbor have been used, which is traditional nonparametric statistical classifier. Set of different experiments on binary images regular, opening image, and closing image of 200 images Batik from some batik literature published by Dinas Perindustrian, Perdagangan, dan Koperasi DIY have been done and variety of results have been presented. The results showed that the best classification result obtained from Hu Moment Invariants feature description.

Classification of remote sensed data using Artificial Bee Colony algorithm

The present study employs the traditional swarm intelligence technique in the classification of satellite data since the traditional statistical classification technique shows limited success in classifying remote sensing data. The traditional statistical classifiers examine only the spectral variance ignoring the spatial distribution of the pixels corresponding to the land cover classes and correlation between various bands. The Artificial Bee Colony (ABC) algorithm based upon swarm intelligence which is used to characterise spatial variations within imagery as a means of extracting information forms the basis of object recognition and classification in several domains avoiding the issues related to band correlation. The results indicate that ABC algorithm shows an improvement of 5% overall classification accuracy at 6 classes over the traditional Maximum Likelihood Classifier (MLC) and Artificial Neural Network (ANN) and 3% against support vector machine.

Sex estimation from the tarsal bones in a Portuguese sample: a machine learning approach.

Sex estimation is extremely important in the analysis of human remains as many of the subsequent biological parameters are sex specific (e.g., age at death, stature, and ancestry). When dealing with incomplete or fragmented remains, metric analysis of the tarsal bones of the feet has proven valuable. In this study, the utility of 18 width, length, and height tarsal measurements were assessed for sex-related variation in a Portuguese sample. A total of 300 males and females from the Coimbra Identified Skeletal Collection were used to develop sex prediction models based on statistical and machine learning algorithm such as discriminant function analysis, logistic regression, classification trees, and artificial neural networks. All models were evaluated using 10-fold cross-validation and an independent test sample composed of 60 males and females from the Identified Skeletal Collection of the 21st Century. Results showed that tarsal bone sex-related variation can be easily captured with a high degree of repeatability. A simple tree-based multivariate algorithm involving measurements from the calcaneus, talus, first and third cuneiforms, and cuboid resulted in 88.3% correct sex estimation both on training and independent test sets. Traditional statistical classifiers such as the discriminant function analysis were outperformed by machine learning techniques. Results obtained show that machine learning algorithm are an important tool the forensic practitioners should consider when developing new standards for sex estimation.

Learning predictors for flash memory endurance: a comparative study of alternative classification methods

Flash memory's ability to be programmed multiple times is called its endurance. Beyond being able to give more accurate chip specifications, more precise knowledge of endurance would permit manufacturers to use flash chips more effectively. Rather than physical testing to determine chip endurance, which is impractical because it takes days and destroys an area of the chip under test, this research seeks to predict whether chips will meet chosen endurance criteria. Timing data relating to erasure and programming operations is gathered as the basis for modelling. The purpose of this paper is to determine which methods can be used on this data to accurately and efficiently predict endurance. Traditional statistical classification methods, support vector machines and genetic programming are compared. Cross-validating on common datasets, the classification methods are evaluated for applicability, accuracy and efficiency and their respective advantages and disadvantages are quantified.

Study on Remote Sensing Image Auto-Identify Classification by Used of Object-Oriented Technology

In dealing with high-resolution remote sensing image auto-identify classification, the traditional pixel-based and spectral statistical characteristics classification technology or method has some insurmountable difficulties. In this paper, object-oriented image analysis method is by application, the auto-identify classification rules are set up based the different remote sensing image characteristics that included such as spectral, texture, scale and so on. As a case study, a petroleum reserve base auto-identify classification is selected as an example and the target is identified, in a better effective result by applications of the object-oriented method. The result appraising analysis indicates that object-oriented classification method to identify automatically high-resolution remote sensing images pattern object can get a high precision. The method of object-oriented has a widely potential application for remote sensing image automatic-identify classification in times to come.

Non-Uniform Random Feature Selection and Kernel Density Scoring With SVM Based Ensemble Classification for Hyperspectral Image Analysis

Traditional statistical classification approaches often fail to yield adequate results with Hyperspectral imagery (HSI) because of the high dimensional nature of the data, multimodal class distribution and limited ground truth samples for training. Over the last decade, Support Vector Machines (SVMs) and Multi-Classifier Systems (MCS) have become popular tools for HSI analysis. Random Feature Selection (RFS) for MCS is a popular approach to produce higher classification accuracies. In this study, we present a Non-Uniform Random Feature Selection (NU-RFS) within a MCS framework using SVM as the base classifier. We propose a method to fuse the output of individual classifiers using scores derived from kernel density estimation. This study demonstrates the improvement in classification accuracies by comparing the proposed approach to conventional analysis algorithms and by assessing the sensitivity of the proposed approach to the number of training samples. These results are compared with that of uniform RFS and regular SVM classifiers. We demonstrate the superiority of Non-Uniform based RFS system with respect to overall accuracy, user accuracies, producer accuracies and sensitivity to number of training samples.

Post Classification Using Cellular Automata for Landsat Images

The research presented in this paper aims at improving the accuracy of land-use maps produced from classification of Landsat images of mega cities in developing countries. In other words, the main objective of this paper is to find a suitable post classification technique that gives optimum results for Landsat images of mega cities in developing countries. To reach our goal, the paper presents a classification of two TM-Landsat sub scenes using a traditional statistical classifier (Maximum Likelihood) into four land cover classes (vegetation-water-Desert-Urban); then the accuracy assessment for the produced land-cover map will be calculated. Following to this step, three post processing techniques- Majority Filter, Probability label Relaxation (PLR), and Cellular Automata (CA) - will be applied in order to improve the accuracy of the previously produced land cover map. Finally, the same accuracy assessment measurements will be calculated for the two land-cover maps produced by each of the above post classification techniques. Initial results will show that CA outperformed the other techniques. In this paper we propose a methodology to implement a satellite image post classification Algorithm with cellular Automata.

Integration of geographic information system and RADARSAT synthetic aperture radar data using a self-organizing map network as compensation for real-time ground data in automatic image classification

The paper presents results of using advanced techniques such as Self-Organizing feature Map (SOM) to incorporate a GIS data layer to compensate for the limited amount of real-time ground-truth data available for land-use and land-cover mapping in wet-season conditions in Bangladesh based on multi-temporal RADARSAT-1 SAR images. The experimental results were compared with those of traditional statistical classifiers such as Maximum Likelihood, Mahalanobis Distance, and Minimum Distance, which are not suitable for incorporating low-level GIS data in the image classification process. The performances of the classifiers were evaluated in terms of the classification accuracy with respect to the collected real-time ground truth data. The SOM neural network provided the highest overall accuracy when a GIS layer of land type classification with respect to the depth and duration of regular flooding was used in the network. Using this method, the overall accuracy was around 15% higher than the previously mentioned traditional classifiers at 79.6% where the training data covered only 0.53% of the total image. It also achieved higher accuracies for more classes in comparison to the other classifiers.

An Automated Artificial Neural Network System for Land Use/Land Cover Classification from Landsat TM Imagery

This paper focuses on an automated ANN classification system consisting of two modules: an unsupervised Kohonen’s Self-Organizing Mapping (SOM) neural network module, and a supervised Multilayer Perceptron (MLP) neural network module using the Backpropagation (BP) training algorithm. Two training algorithms were provided for the SOM network module: the standard SOM, and a refined SOM learning algorithm which incorporated Simulated Annealing (SA). The ability of our automated ANN system to perform Land-Use/Land-Cover (LU/LC) classifications of a Landsat Thematic Mapper (TM) image was tested using a supervised MLP network, an unsupervised SOM network, and a combination of SOM with SA network. Our case study demonstrated that the ANN classification system fulfilled the tasks of network training pattern creation, network training, and network generalization. The results from the three networks were assessed via a comparison with reference data derived from the high spatial resolution Digital Colour Infrared (CIR) Digital Orthophoto Quarter Quad (DOQQ) data. The supervised MLP network obtained the most accurate classification accuracy as compared to the two unsupervised SOM networks. Additionally, the classification performance of the refined SOM network was found to be significantly better than that of the standard SOM network essentially due to the incorporation of SA. This is mainly due to the SA-assisted classification utilizing the scheduling cooling scheme. It is concluded that our automated ANN classification system can be utilized for LU/LC applications and will be particularly useful when traditional statistical classification methods are not suitable due to a statistically abnormal distribution of the input data.