Null Entries Research Articles

Dichotomization of Multilevel Variables to Detect Hidden Associations

A test of independence is commonly used to determine differences (or associations) between samples in a nominal level measurement. Fisher’s exact test and Chi-square test are two of the most widely applied tests of independence used in the data analyses in different areas such as information technologies, biostatistics, psychology and health sciences. In some cases, contingency tables with null entries (also called random zeros) arise, particularly if the number of samples is small, and the variables analyzed are multilevel. This situation becomes a problem because if one or more entries in a contingency table are zero or have small values, then the tests of independence produce unreliable results. In this paper, we propose a method to address that issue. The method merges one or more levels of the variables analyzed to create contingency tables with only one degree of freedom, avoiding applying a test of independence on contingency tables with random zeros. The source code (Python) of the method is publicly available for use. The results obtained using our method give a complete panorama of the associations between the variables of a data set. To show the effectiveness of our approach to find dependencies between variables, we use four data sets publicly available on the Internet.

Detecting Cyberbullying in Roman Urdu Language Using Natural Language Processing Techniques

Nowadays, social media platforms are the primary source of public communication and information. Social media platforms have become an integral part of our daily lives, and their user base is rapidly expanding as access is extended to more remote locations. Pakistan has around 71.70 million social media users that utilize Roman Urdu to communicate. With these improvements and the increasing number of users, there has been an increase in digital bullying, often known as cyberbullying. This research focuses on social media users who use Roman Urdu (Urdu language written in the English alphabet) to communicate. In this research, we explored the topic of cyberbullying actions on the Twitter platform, where users employ Roman Urdu as a medium of communication. To our knowledge, this is one of the very few studies that address cyberbullying behavior in Roman Urdu. Our proposed study aims to identify a suitable model for classifying cyberbullying behavior in Roman Urdu. To begin, the dataset was designed by extracting data from twitter using twitter's API. The targeted data was extracted using keywords based on Roman Urdu. The data was then annotated as bully and not-bully. After that, the dataset has been pre-processed to reduce noise, which includes punctuation, stop words, null entries, and duplication removal. Following that, features are extracted using two different methods, Count-Vectorizer and TF-IDF Vectorizer, and a set of ten different learning algorithms including SVM, MLP, and KNN was applied to both types of extracted features based on supervised learning. Support Vector Machine (SVM) performed the best out of the implemented algorithms by both combinations, with 97.8 percent when implemented over the TF-IDF features and 93.4 percent when implemented over the CV features. The proposed mechanism could be helpful for online social apps and chat rooms for the better detection and designing of bully word filters, making safer cyberspace for end users.

Association of an immunosuppressive macrophage transcriptional signature derived from peripheral blood macrophages with gastric cancer prognosis.

354 Background: Gastric cancer (GC) is the fourth most common cause of cancer death worldwide. Studies have shown an association of higher macrophage density in the tumor microenvironment with worse outcomes in GC. The presence of M2 specific gene signatures in the peripheral blood of cancer patients are associated with poor prognosis. However, it is unknown whether these immunosuppressive M2 markers are prognostic when found in the primary GC tumors. Methods: Based on previously published differential gene expression profiles that distinguish M1 versus M2 macrophages in the peripheral blood, we used the UCSC Xena webtool to analyze the TCGA Stomach Cancer cohort for tumor-associated immune signatures for their prognostic value. We excluded null entries, normal tissue, and entries with incomplete data. We investigated both M1- and M2-macrophage signatures (M1 signature: CCR7, IL2RA, CXCL11, CCL19, CXCL10, PLA1A, PTX3 and M2 signature: MRC1, MS4A4A, CD36, CCL13, CCL18, CCL23, SLC38A6, FGL2, FN1, MAF); as well as a T cell cytolytic signature defined by genes with cytolytic functions ( GZMA, GZMB, GZMH, GZMM, PRF1). We analyzed the impact of transcriptional gene expression with relation to stage, histological subtype, overall survival (OS), and disease free interval (DFI). The UCSC Xena webtool uses log-rank test to compare Kaplan-Meier curves. Results: A total of 341 entries were analyzed with age ranging from 30-90 and 39.4% female. Increased CD168 expression was associated with increased stage of GC (p = 0.017). Higher expression of the M2-macrophage signature was associated with diffuse type adenocarcinoma and mucinous type intestinal adenocarcinoma, compared to tubular and papillary type intestinal adenocarcinoma (p < 0.001). In addition, higher M2-macrophage signature expression was associated with late stage (Stage II or above) compared to early stage (stage I) (p = 0.022). A low M2-macrophage signature was associated with improved 5-year OS (p = 0.029), and a significantly improved 5-year DFI (p = 0.042). CD8A, cytolytic T cell signature, and M1 macrophage signature were not associated with stage of disease or survival. Conclusions: A transcriptional signature derived from peripheral blood immunosuppressive M2 macrophage phenotype is found in GC primary tumors and associated with higher pathologic staging and with worse outcomes. These M2 defining molecular biomarkers of GC has the potential to serve as targets of novel immune-therapeutic strategies.

Pakistani News Classification Based on Headlines

In information dissemination, media plays a vital role, a huge amount of information is spread quickly through social media and news platforms. The online news outlet and frequency of the huge number of headlines creation demand automatic classification tools. In this overwhelming domain, different countries explored classification algorithms for their news. However, very limited research is done on Pakistani news headlines in terms of classification. There is also a lack of a Pakistani news headlines benchmark dataset. This research is intended to find a suitable model to classify Pakistani news headlines automatically. Moreover, in this study, we designed a benchmark NCE-2D (News Classification and Emotion Detection Dataset) of 2429 Pakistani news headlines extracted from different news websites using ParseHub, which includes five types of categories. First, to minimize the noise from the dataset is undergoing some pre-processing steps including punctuation, stopwords, null entries, and duplication removal. In the next step, extract the feature by two different methods Count-Vectorizer and TF-IDF Vectorizer, and then apply a set of 10 different learning algorithms for both types of features extracted. The best performing algorithms for Pakistani news headlines out of the set of implemented algorithms by both combinations included Support Vector Machine (SVM) when implemented over the TF-IDF features and Multi-Layer Perception (MLP) when implemented over the CV features. Based on the results, both these algorithms are compared to choose the most suitable from them. The combination of SVM is investigated most appropriate classifier with 82.16% accuracy than 81.75% accuracy of MLP.

Emotion Classification through Product Consumer Reviews

Currently, the world is becoming digitalized. Ecommerce is enamoring strength in this digitalized world through the accessibility of products reachable for customers. The recent work in the field of emotion classification based on consumer feedback reviews is gaining popularity worldwide. However, the research lacks to provide better accuracy in terms of pre-processing and feature extraction. Since, to determine consumer emotions in terms of their choices, feature extraction and pre-processing plays an important role to enhance the accuracy rate to determine their choices for product marketability and to identify consumer behavior in selecting product features resulting in increased product demands. The review of related literature shows that the products accuracy rate can be improved though refining the dataset by executing proper pre-processing techniques like removing null values and meaning less values from the dataset to make it more accurate and examining its feature extraction by applying different algorithms which required a lot of pre-processing of reviewed data to make the features more valuable. This research provides an analysis of cars, hotels and mobile datasets and studies sentiment classification with different machine learning Approaches. First of all, to minimize the noise of the dataset was undergoing with pre-processing steps including punctuation, stop words, null entries and duplication removal. In the next step, extract the feature by two different methods: Count Vectorizer and bag of words. After that bidirectional encoder representation from transformers (BERT) algorithm were applied on three datasets to predict the results. By applying BERT Classifier on cars dataset 98.5% accuracy were founded, by applying same algorithm on hotels dataset 98.3% accuracy were founded, and for Mobile dataset 98.7% accuracy were founded.

Hybrid Method of ACA and Interpolation-based Algorithms for Analysis of Eddy Current Nondestructive Evaluations

ABSTRACT In this article, the hybridization of adaptive cross approximation (ACA) algorithm and interpolation-based separation of the kernel function is proposed to accelerate solving the matrix equations resulted in the boundary element method (BEM) for 3D arbitrary-shaped eddy current nondestructive evaluation problems. The hybrid method combines the advantages of both the ACA algorithm and the interpolation-based methods, and resolves the shortcoming of pure ACA method, when modeling the planar eddy current nondestructive evaluation problems, that it cannot compress the null entries the BEM generated when the testing and basis patches are co-planar. In the proposed method, the submatrices associated with the null entries are compressed by the interpolation-based method, while the others are compressed by the ACA algorithm. Several benchmarks are shown to demonstrate both the robustness and efficiency of the proposed fast and general solver.

On the characterization of the space of derivations in evolution algebras

We study the space of derivations for some finite-dimensional evolution algebras, depending on the twin partition of an associated directed graph. For evolution algebras with a twin-free associated graph, we prove that the space of derivations is zero. For the remaining families of evolution algebras, we obtain sufficient conditions under which the study of such a space can be simplified. We accomplish this task by identifying the null entries of the respective derivation matrix. Our results suggest how strongly the associated graph’s structure impacts in the characterization of derivations for a given evolution algebra. Therefore, our approach constitutes an alternative to the recent developments in the research of this subject. As an illustration of the applicability of our results, we provide some examples and we exhibit the classification of the derivations for non-degenerate irreducible three-dimensional evolution algebras.

Complex Activity Recognition Using Acceleration, Vital Sign, and Location Data

Complex activity recognition is a valuable issue in mobile and wearable computing. Since complex activities are strongly relevant to users’ locations, location data can be used in complex activity recognition. Current activity recognition methods utilizing location data either do not make the most of location data or have a “cold-start” problem. In this paper, we propose CAROLINA (complex activity recognition using acceleration, vital sign, and location data) to recognize complex activities from wearable sensors. CAROLINA divides a map into grid cells and builds a grid cell-activity matrix by integrating check-in and ATUS (American Time Use Survey) datasets. Since some grid cells do not have check-in records, there are some null entries in the matrix, and all the entries are reconstructed by matrix factorization. Meanwhile, a POI (point of interest) category dataset is exploited to build a grid cell-grid cell similarity matrix, which is used to reduce the loss of matrix factorization. The results of experiments show that CAROLINA can effectively utilize location data to improve complex activity recognition performance.

Exactly Solvable Record Model for Rainfall.

Daily precipitation time series are composed of null entries corresponding to dry days and nonzero entries that describe the rainfall amounts on wet days. Assuming that wet days follow a Bernoulli process with success probability p, we show that the presence of dry days induces negative correlations between record-breaking precipitation events. The resulting nonmonotonic behavior of the Fano factor of the record counting process is recovered in empirical data. We derive the full probability distribution P(R,n) of the number of records R_{n} up to time n, and show that for large n, it converges to a Poisson distribution with parameter ln(pn). We also study in detail the joint limit p→0, n→∞, which yields a random record model in continuous time t=pn.

Risk-Index of Colorectal Cancer to Triage for Screening

Colorectal cancer (CRC) is the unregulated growth of one or more adenomatous polyps (most frequently adenocarcinomas) occurring in the colon and/or the rectum. Of all new cancer cases in the US, 8.0% are colorectal. Colorectal cancer claims 8.4% of all cancer-deaths and thus has above-average mortality vs. other cancers. To detect CRC at early stages, the United States Preventative Services Task Force (USPSTF) recommends screening by colonoscopy or sigmoidoscopy for ages 50-75. However, the high false positive rates of these procedures lead to a lot of unnecessary screening. The goal of this study is to develop an artificial neural network (ANN) for colorectal cancer risk prediction usable for triaging people for screening based on their personal health data. The data to train and validate the ANN is the set of 1997-2016 responses (excepting 2004) to the National Health Interview Survey (NHIS) personal health questionnaire, in which colon and/or rectal cancer occurring 4 years or less from the survey date is counted as one instance of CRC. Respondents with 1 or more null entries were discarded. Using 70% of the NHIS data, training uses gradient descent in parameter space to adjust the parameters of a softmax-function and minimize a logistic cost function (regression). The extent to which the parameters are adjusted is calculated using the chain rule of calculus (backpropagation). The validation tests the regression-fitted function upon the remaining 30% of the data. As a binary test, sensitivity (TPR), specificity (SPC), and positive/negative predictive values (PPV/NPV) are calculated and compared with the USPSTF. As a trinary test, three levels of risk-stratification are defined by requiring that no more than 1% of CRC/non-CRC cases be misclassified as low/high-risk. If a person of medium risk having/not having CRC is counted as half of a false negative/positive (e.g., annual/biennial screenings for those of high/medium risk), the trinary test of risk-stratification also has an associated TPR, SPC, PPV, and NPV. As a binary test, the ANN has sensitivity of 0.7, specificity of 0.7, PPV of 0.09, and NPV of 0., all (except for NPV) exceeding the USPSTF guidelines and independent of tumorous advancement. As a trinary test, lowered SPC and PPV are exchanged for higher TPR and NPV.Abstract MO_30_2748; Table 1Performance of USPSTF (entire data set) compared to the ANN (validation-portion of data set, only).CRC, # screenedCRC, # not screenedNo CRC, # screenedNo CRC, # not screenedANN (bi.)517255106,864235,029ANN (tri.)497145142,603.5207,004USPSTF762490232,545297,065TPRSPCPPVNPVANN (bi.)0.6700.6870.009450.997ANN (tri.)0.7740.4080.003470.9993USPSTF0.6090.4390.003270.998 Open table in a new tab CRC risk calculated by ANN from personal health data is noninvasive, insensitive to tumorous advancement, & outperforms USPSTF screening guidelines as a binary and trinary test. In addition, the ANN offers the prospect of CRC risk assessment in real time and on the world map.

Matrix decompositions using sub-Gaussian random matrices

Abstract In recent years, several algorithms which approximate matrix decomposition have been developed. These algorithms are based on metric conservation features for linear spaces of random projection types. We present a new algorithm, which achieves with high probability a rank-$r$ singular value decomposition (SVD) approximation of an $n \times n$ matrix and derive an error bound that does not depend on the first $r$ singular values. Although the algorithm has an asymptotic complexity similar to state-of-the-art algorithms and the proven error bound is not as tight as the state-of-the-art bound, experiments show that the proposed algorithm is faster in practice while providing the same error rates as those of the state-of-the-art algorithms. We also show that an i.i.d. sub-Gaussian matrix with large probability of having null entries is metric conserving. This result is used in the SVD approximation algorithm, as well as to improve the performance of a previously proposed approximated LU decomposition algorithm.

Variable selection in multivariate linear models with high-dimensional covariance matrix estimation

In this paper, we propose a novel variable selection approach in the framework of multivariate linear models taking into account the dependence that may exist between the responses. It consists in estimating beforehand the covariance matrix Σ of the responses and to plug this estimator in a Lasso criterion, in order to obtain a sparse estimator of the coefficient matrix. The properties of our approach are investigated both from a theoretical and a numerical point of view. More precisely, we give general conditions that the estimators of the covariance matrix and its inverse have to satisfy in order to recover the positions of the null and non null entries of the coefficient matrix when the size of Σ is not fixed and can tend to infinity. We prove that these conditions are satisfied in the particular case of some Toeplitz matrices. Our approach is implemented in the R package MultiVarSel available from the Comprehensive R Archive Network (CRAN) and is very attractive since it benefits from a low computational load. We also assess the performance of our methodology using synthetic data and compare it with alternative approaches. Our numerical experiments show that including the estimation of the covariance matrix in the Lasso criterion dramatically improves the variable selection performance in many cases.

Smallest singular value of sparse random matrices

We extend probability estimates on the smallest singular value of random matrices with independent entries to a class of sparse random matrices. We show that one can relax a previously used condition of uniform boundedness of the variances from below. This allows us to consider matrices with null entries or, more generally, with entries having small variances. Our results do not assume identical distribution of the entries of a random matrix and help to clarify the role of the variances of the entries. We also show that it is enough to require boundedness from above of the r-th moment, r > 2, of the corresponding entries. AMS 2010 Classification: 46B06, 60B20, 15B52

Metabolite profiling in human urine by LC–MS/MS: Method optimization and application for glucuronides from dextromethorphan metabolism

Analysis of human urine for specific compounds or metabolites is an established method for biomonitoring occupational or environmental exposures. Modern liquid chromatography–tandem mass spectrometry is not limited to single compounds but can simultaneously analyze whole classes of urine constituents with both high sensitivity and specificity. Individual differences in the composition of urine are very large in humans, which raises a number of problems that are not encountered in animal experimentation. In this report, we investigated whether analysis of glucuronides as a class could reflect differences between human individuals regarding the polymorphic activity of the cytochrome P450 enzyme CYP2D6. From a group of 152 students that had been classified for CYP2D6 activity, urine of 12 “poor metabolizers” and 35 “extensive metabolizers” was collected 90 min after ingestion of 10 mg of the antitussive drug dextromethorphan (DEX) and analyzed for glucuronides. Methods development included the following aspects: adjustment of urine samples to equal creatinine concentration to avoid differences between samples in retention times and ion suppression; on-line enrichment of low-level analytes by column switching; precursor ion scan vs. theoretical multiple reaction monitoring; use of quality control samples to check for reproducibility in large sample series; peak extraction and handling of null entries to build the data matrix; logarithmic data transformation and different scaling procedures; principal component analysis (PCA) vs. discriminant analysis. Our results show that an optimized procedure not only identified the known DEX metabolites as predictors of CYP2D6-specific metabolic pathways but also indicated the presence of additional, so far unknown path-specific glucuronide metabolites. We conclude that metabolite profiling of urine and other biofluids by modern mass spectrometric methodology may help characterize individual differences and become useful in drug development and personalized pharmacotherapy.

More on deviation from bi-maximal neutrino mixing

We study the case of UTl presenting the exact bi-maximal mixing form with Uν inducing the deviation from the bi-maximal mixing in the final form of the Pontecorvo–Maki–Nakagawa–Sakata neutrino mixing, UPMNS = UTlUν. We will show that such a possibility will lead to a democratic texture for the charged lepton mass matrix and to a neutrino mass matrix with four null entries.

Reflection K-matrices for 19-vertex models

We derive and classify all regular solutions of the boundary Yang-Baxter equation for 19-vertex models known as Zamolodchikov-Fateev or A 1 (1) model, Izergin-Korepin or A 2 (2) model, sl(2|1) model and the osp(2|1) model. We find that there is a general solution for A 1 (10) and sl(2|1) models. In both models it is a complete K-matrix with three free parameters. For the A 2 (2) and os(2|1) models we find three general solutions, being two complete reflection K-matrices solutions and one incomplete reflection K-matrix solution with some null entries. In both models these solutions have two free parameters. Integrable spin-1 Hamiltonians with general boundary interactions are also presented. Several reduced solutions from these general solutions are presented in the appendices.

Managing Null Entries in Pairwise Comparisons

This paper shows how to manage null entries in pairwise comparisons matrices. Although assessments can be imprecise, since subjective criteria are involved, the classical pairwise comparisons theory expects all of them to be available. In practice, some experts may not be able (or available) to provide all assessments. Therefore managing null entries is a necessary extension of the pairwise comparisons method. It is shown that certain null entries can be recovered on the basis of the transitivity property which each pairwise comparisons matrix is expected to satisfy.