scholarly journals A regularized logistic regression based model for supervised learning

2020 ◽  
Vol 14 ◽  
pp. 174830262097153
Author(s):  
Carlos Brito-Pacheco ◽  
Carlos Brito-Loeza ◽  
Anabel Martin-Gonzalez
Keyword(s):  
Logistic Regression ◽  
Mathematical Analysis ◽  
Logistic Model ◽  
Supervised Classification ◽  
State Of The Art ◽  
Logistic Models ◽  
Weight Vector ◽  
Classification Problem ◽  
Gradient Vector ◽  
Continuous Setting

In this work, we introduce a new regularized logistic model for the supervised classification problem. Current logistic models have become the preferred tools for supervised classification in many situations. They mostly use either L1 or L2 regularization of the weight vector of parameters. Here we take a different approach by applying regularization not to the weight vector but to the gradient vector of the function representing the separating hyper-surface. We present the mathematical analysis of the model in its continuous setting and provide experimental evidence to show that the new model is competitive with state of the art models.

scholarly journals Comparison of Logistic Regression and Linear Regression in Modeling Percentage Data

2001 ◽  
Vol 67 (5) ◽  
pp. 2129-2135 ◽  
Author(s):  
Lihui Zhao ◽  
Yuhuan Chen ◽  
Donald W. Schaffner
Keyword(s):  
Logistic Regression ◽  
Linear Regression ◽  
Logistic Model ◽  
Goodness Of Fit ◽  
Logistic Models ◽  
Food Microbiology ◽  
Data Sets ◽  
Data Set ◽  
Predicted Values ◽  
Percentage Data

ABSTRACT Percentage is widely used to describe different results in food microbiology, e.g., probability of microbial growth, percent inactivated, and percent of positive samples. Four sets of percentage data, percent-growth-positive, germination extent, probability for one cell to grow, and maximum fraction of positive tubes, were obtained from our own experiments and the literature. These data were modeled using linear and logistic regression. Five methods were used to compare the goodness of fit of the two models: percentage of predictions closer to observations, range of the differences (predicted value minus observed value), deviation of the model, linear regression between the observed and predicted values, and bias and accuracy factors. Logistic regression was a better predictor of at least 78% of the observations in all four data sets. In all cases, the deviation of logistic models was much smaller. The linear correlation between observations and logistic predictions was always stronger. Validation (accomplished using part of one data set) also demonstrated that the logistic model was more accurate in predicting new data points. Bias and accuracy factors were found to be less informative when evaluating models developed for percentage data, since neither of these indices can compare predictions at zero. Model simplification for the logistic model was demonstrated with one data set. The simplified model was as powerful in making predictions as the full linear model, and it also gave clearer insight in determining the key experimental factors.

scholarly journals Quantification of rhBMP2 in bioactive bone materials

2019 ◽  
Vol 7 (1) ◽  
pp. 71-75
Author(s):  
Huan Lian ◽  
Han Wang ◽  
Qianqian Han ◽  
Chunren Wang
Keyword(s):  
Logistic Regression ◽  
Regression Models ◽  
Logistic Model ◽  
Transforming Growth Factor ◽  
Logistic Models ◽  
Transforming Growth Factor Β ◽  
C2c12 Cells ◽  
Logistic Regression Models ◽  
Alp Activity

Abstract Bone morphogenetic protein (BMP), belongs to transforming growth factor-β (TGF-β) superfamily except BMP-1. Implanting BMP into muscular tissues induces ectopic bone formation at the site of implantation, which provides opportunity for the treatment of bone defects. Recombinant human BMP-2 (rhBMP-2) has been used clinically, but the lack of standard methods for quantifying rhBMP-2 biological activity greatly hindered the progress of commercialization. In this article, we describe an in vitro rhBMP-2 quantification method, as well as the data analyzation pipeline through logistic regression in RStudio. Previous studies indicated that alkaline phosphatase (ALP) activity of C2C12 cells was significantly increased when exposed to rhBMP-2, and showed dose-dependent effects in a certain concentration range of rhBMP-2. Thus, we chose to quantify ALP activity as an indicator of rhBMP-2 bioactivity in vitro. A sigmoid relationship between the ALP activity and concentration of rhBMP-2 was discovered. However, there are tons of regression models for such a non-linear relationship. It has always been a major concern for researchers to choose a proper model that not only fit data accurately, but also have parameters representing practical meanings. Therefore, to fit our rhBMP-2 quantification data, we applied two logistic regression models, three-parameter log-logistic model and four-parameter log-logistic model. The four-parameter log-logistic model (adj-R2 > 0.98) fits better than three-parameter log-logistic model (adj-R2 > 0.75) for the sigmoid curves. Overall, our results indicate rhBMP-2 quantification in vitro can be accomplished by detecting ALP activity and fitting four-parameter log-logistic model. Furthermore, we also provide a highly adaptable R script for any additional logistic models.

scholarly journals Structural Health Monitoring of 2D Plane Structures

2021 ◽  
Vol 11 (5) ◽  
pp. 2000
Author(s):  
Behnam Mobaraki ◽  
Haiying Ma ◽  
Jose Antonio Lozano Galant ◽  
Jose Turmo
Keyword(s):  
Mechanical Properties ◽  
Health Monitoring ◽  
Mathematical Analysis ◽  
Stiffness Matrix ◽  
State Of The Art ◽  
System Of Equations ◽  
Matrix System ◽  
Structural System ◽  
Detailed Mathematical Analysis ◽  
Structural System Identification

This paper presents the application of the observability technique for the structural system identification of 2D models. Unlike previous applications of this method, unknown variables appear both in the numerator and the denominator of the stiffness matrix system, making the problem non-linear and impossible to solve. To fill this gap, new changes in variables are proposed to linearize the system of equations. In addition, to illustrate the application of the proposed procedure into the observability method, a detailed mathematical analysis is presented. Finally, to validate the applicability of the method, the mechanical properties of a state-of-the-art plate are numerically determined.

scholarly journals The use of the nonlinear models in the growth of pears of ‘Shinseiki’ cultivar

2017 ◽  
Vol 48 (1) ◽  
Author(s):  
Thais Destefani Ribeiro ◽  
Taciana Villela Savian ◽  
Tales Jesus Fernandes ◽  
Joel Augusto Muniz
Keyword(s):  
Logistic Model ◽  
Least Squares Method ◽  
Nonlinear Models ◽  
Logistic Models ◽  
Information Criterion ◽  
Fruit Weight ◽  
Coefficient Of Determination ◽  
Pear Fruit ◽  
Asian Pear ◽  
Residual Correlations

ABSTRACT: The goal of this study was to elucidate the growth and development of the Asian pear fruit, on the grounds of length, diameter and fresh weight determined over time, using the non-linear Gompertz and Logistic models. The specifications of the models were assessed utilizing the R statistical software, via the least squares method and iterative Gauss-Newton process (DRAPER & SMITH, 2014). The residual standard deviation, adjusted coefficient of determination and the Akaike information criterion were used to compare the models. The residual correlations, observed in the data for length and diameter, were modeled using the second-order regression process to render the residuals independent. The logistic model was highly suitable in demonstrating the data, revealing the Asian pear fruit growth to be sigmoid in shape, showing remarkable development for three variables. It showed an average of up to 125 days for length and diameter and 140 days for fresh fruit weight, with values of 72mm length, 80mm diameter and 224g heavy fat.

scholarly journals A Mating Selection Based on Modified Strengthened Dominance Relation for NSGA-III

Mathematics ◽  
2021 ◽  
Vol 9 (22) ◽  
pp. 2837
Author(s):  
Saykat Dutta ◽  
Sri Srinivasa Raju M ◽  
Rammohan Mallipeddi ◽  
Kedar Nath Das ◽  
Dong-Gyu Lee
Keyword(s):  
Pareto Front ◽  
State Of The Art ◽  
Weight Vector ◽  
Dominance Relation ◽  
Additional Parameter ◽  
Pareto Dominance ◽  
Environmental Selection ◽  
Dominance Relationships ◽  
Test Suites ◽  
Selection Of

In multi/many-objective evolutionary algorithms (MOEAs), to alleviate the degraded convergence pressure of Pareto dominance with the increase in the number of objectives, numerous modified dominance relationships were proposed. Recently, the strengthened dominance relation (SDR) has been proposed, where the dominance area of a solution is determined by convergence degree and niche size (θ¯). Later, in controlled SDR (CSDR), θ¯ and an additional parameter (k) associated with the convergence degree are dynamically adjusted depending on the iteration count. Depending on the problem characteristics and the distribution of the current population, different situations require different values of k, rendering the linear reduction of k based on the generation count ineffective. This is because a particular value of k is expected to bias the dominance relationship towards a particular region on the Pareto front (PF). In addition, due to the same reason, using SDR or CSDR in the environmental selection cannot preserve the diversity of solutions required to cover the entire PF. Therefore, we propose an MOEA, referred to as NSGA-III*, where (1) a modified SDR (MSDR)-based mating selection with an adaptive ensemble of parameter k would prioritize parents from specific sections of the PF depending on k, and (2) the traditional weight vector and non-dominated sorting-based environmental selection of NSGA-III would protect the solutions corresponding to the entire PF. The performance of NSGA-III* is favourably compared with state-of-the-art MOEAs on DTLZ and WFG test suites with up to 10 objectives.

scholarly journals An Analysis of State-of-the-art Activation Functions For Supervised Deep Neural Network

2021 ◽  
Author(s):  
Anh Nguyen ◽  
Khoa Pham ◽  
Dat Ngo ◽  
Thanh Ngo ◽  
Lam Pham
Keyword(s):  
Neural Network ◽  
Supervised Classification ◽  
Deep Neural Network ◽  
State Of The Art ◽  
Network Architectures ◽  
Activation Functions ◽  
Scene Classification ◽  
Learning Network ◽  
Deep Learning Network

This paper provides an analysis of state-of-the-art activation functions with respect to supervised classification of deep neural network. These activation functions comprise of Rectified Linear Units (ReLU), Exponential Linear Unit (ELU), Scaled Exponential Linear Unit (SELU), Gaussian Error Linear Unit (GELU), and the Inverse Square Root Linear Unit (ISRLU). To evaluate, experiments over two deep learning network architectures integrating these activation functions are conducted. The first model, basing on Multilayer Perceptron (MLP), is evaluated with MNIST dataset to perform these activation functions.Meanwhile, the second model, likely VGGish-based architecture, is applied for Acoustic Scene Classification (ASC) Task 1A in DCASE 2018 challenge, thus evaluate whether these activation functions work well in different datasets as well as different network architectures.

scholarly journals nMNSD—A Spiking Neuron-Based Classifier That Combines Weight-Adjustment and Delay-Shift

2021 ◽  
Vol 15 ◽  
Author(s):  
Gianluca Susi ◽  
Luis F. Antón-Toro ◽  
Fernando Maestú ◽  
Ernesto Pereda ◽  
Claudio Mirasso
Keyword(s):  
State Of The Art ◽  
Classification Problem ◽  
Mathematical Framework ◽  
Recognition Mechanism ◽  
Specific Input ◽  
Biological Learning ◽  
Time And Energy ◽  
Spike Latency ◽  
Spike Sequence

The recent “multi-neuronal spike sequence detector” (MNSD) architecture integrates the weight- and delay-adjustment methods by combining heterosynaptic plasticity with the neurocomputational feature spike latency, representing a new opportunity to understand the mechanisms underlying biological learning. Unfortunately, the range of problems to which this topology can be applied is limited because of the low cardinality of the parallel spike trains that it can process, and the lack of a visualization mechanism to understand its internal operation. We present here the nMNSD structure, which is a generalization of the MNSD to any number of inputs. The mathematical framework of the structure is introduced, together with the “trapezoid method,” that is a reduced method to analyze the recognition mechanism operated by the nMNSD in response to a specific input parallel spike train. We apply the nMNSD to a classification problem previously faced with the classical MNSD from the same authors, showing the new possibilities the nMNSD opens, with associated improvement in classification performances. Finally, we benchmark the nMNSD on the classification of static inputs (MNIST database) obtaining state-of-the-art accuracies together with advantageous aspects in terms of time- and energy-efficiency if compared to similar classification methods.

scholarly journals A recommendations model with multiaspect awareness and hierarchical user-product attention mechanisms

2020 ◽  
Vol 17 (3) ◽  
pp. 849-865
Author(s):  
Zhongqin Bi ◽  
Shuming Dou ◽  
Zhe Liu ◽  
Yongbin Li
Keyword(s):  
State Of The Art ◽  
Weight Vector ◽  
User Preferences ◽  
The Other ◽  
Attention Networks ◽  
Proposed Model ◽  
Network Methods ◽  
Public Datasets ◽  
Novel Model ◽  
Attention Weight

Neural network methods have been trained to satisfactorily learn user/product representations from textual reviews. A representation can be considered as a multiaspect attention weight vector. However, in several existing methods, it is assumed that the user representation remains unchanged even when the user interacts with products having diverse characteristics, which leads to inaccurate recommendations. To overcome this limitation, this paper proposes a novel model to capture the varying attention of a user for different products by using a multilayer attention framework. First, two individual hierarchical attention networks are used to encode the users and products to learn the user preferences and product characteristics from review texts. Then, we design an attention network to reflect the adaptive change in the user preferences for each aspect of the targeted product in terms of the rating and review. The results of experiments performed on three public datasets demonstrate that the proposed model notably outperforms the other state-of-the-art baselines, thereby validating the effectiveness of the proposed approach.

scholarly journals Selected Robust Logistic Regression Specification for Classification of Multi‑dimensional Functional Data in Presence of Outlier

2018 ◽  
Vol 2 (334) ◽  
Author(s):  
Mirosław Krzyśko ◽  
Łukasz Smaga
Keyword(s):  
Logistic Regression ◽  
Regression Model ◽  
Functional Data ◽  
Logistic Regression Model ◽  
Binary Classification ◽  
Classification Problem ◽  
Classification Rule ◽  
Unknown Parameters ◽  
Explanatory Variables

In this paper, the binary classification problem of multi‑dimensional functional data is considered. To solve this problem a regression technique based on functional logistic regression model is used. This model is re‑expressed as a particular logistic regression model by using the basis expansions of functional coefficients and explanatory variables. Based on re‑expressed model, a classification rule is proposed. To handle with outlying observations, robust methods of estimation of unknown parameters are also considered. Numerical experiments suggest that the proposed methods may behave satisfactory in practice.

Feeling Not Wanted/Loved and Depression: Does Gender Matter?

2021 ◽  
pp. 003329412110628
Author(s):  
Manik Ahuja ◽  
Joy Okoro ◽  
Esther Frimpong ◽  
Riddhi P. Doshi ◽  
Rajvi J. Wani
Keyword(s):  
Logistic Regression ◽  
Regression Analysis ◽  
Parental Attachment ◽  
Add Health ◽  
Logistic Models ◽  
Men And Women ◽  
Future Studies ◽  
National Longitudinal Study ◽  
Family Based ◽  
Family Based Interventions

Depression affects over 17 million American adults yearly and has been identified as the leading cause of disability in people between the ages of 15 and 44 years. There is evidence that feeling neglect or a lack of parental attachment during childhood is associated with depression. One construct that has been overlooked is love from a parent. The purpose of this study was to analyze the association between individuals who felt not wanted/loved during adolescence and lifetime depression and to examine this association by gender. We examined 5114 participants aged 24–32 years at Wave IV of the National Longitudinal Study of Adolescent Health (Add Health) public use dataset. We used logistic regression analysis to determine the association between an individual feeling not wanted/loved by their parent/caregiver prior to age 18, covariates, and lifetime depression. We then stratified by gender and ran logistic models for both men and women separately. Overall, 16.2% ( n = 827) reported lifetime depression diagnosis, while 16.5% of the participants reported feeling not wanted/loved “often,” while 29.8% reported it as “sometimes.” Feeling not wanted/loved “often” was associated with higher odds of depression (AOR = 3.00; 95% CI, 2.45–3.66; p < .001) versus “sometimes” (AOR = 1.59; 95% CI, 1.31–1.90; p < .001). When stratified by gender, feeling not loved/wanted was associated with depression among both men (AOR = 3.70; 95% CI, 2.60–5.25; p < .001) and women (AOR = 2.73; 95% CI, 2.13–3.48; p < .001). Feeling not loved or wanted by a parent/caregiver during adolescence has serious implications, for both men and women. Future studies should further examine this construct and identify family-based interventions that focus on parent/caregiver and child relationships.

Sign in / Sign up

Export Citation Format

Share Document