scholarly journals Investigating the Interaction Between Prosody and Pragmatics Quantitatively: A Case Study of the Chinese Discourse Marker ni zhidao (“You Know”)

2021 ◽  
Vol 12 ◽  
Author(s):  
Yi Shan
Keyword(s):  
Statistical Analysis ◽  
Random Forest ◽  
Classification Model ◽  
Prosodic Features ◽  
Discourse Marker ◽  
Model Based ◽  
Communicative Functions ◽  
Pragmatic Awareness ◽  
Instrumental Methods

This study briefly describes the prosodic and pragmatic characteristics of the discourse marker ni zhidao (“you know”) in spoken Chinese. It mainly explores the interaction between its prosody and pragmatics using instrumental methods. It is the first attempt to use acoustic and statistical analysis to examine the prosodic parameters and prosody-pragmatics interaction of a Chinese discourse marker. The corpus includes 71 interview conversations totaling more than 30 h, in which 490 discourse marker tokens of ni zhidao were found. Ni zhidao mainly fulfilled four broad pragmatic functions of initiating a topic when occurring sentence-initially, of holding the floor when appearing within clauses, of marking coherence when making its presence between clauses, and of projecting attitudes and feelings when showing up sentence-finally. Drawing on the algorithm of random forest in R, the acoustic and statistical analysis of the performance of ni zhidao in these four functions showed that its prosodic features, including duration, tempo, pre-pause, post-pause, F0, and intensity, significantly relate to and thus imply its pragmatic functions, that the interaction between its prosody and pragmatics can be modeled statistically, and that the established pragmatics classification model based on prosody can be utilized to predict the pragmatics of ni zhidao. These findings seem to strengthen the hypothesis that prosodic variables play a role in deciphering the different pragmatic functions of ni zhidao. This study uses prosodic evidence to more objectively reveal not only the part of ni zhidao in dynamically constructing and embodying specific contexts but also its communicative functions and the underlying meta-pragmatic awareness behind it. This study breaks through the limitations of traditional discourse marker research, which mainly relies on context and discourse characteristics for subjective reasoning.

scholarly journals Classification and prediction of Mycobacterium Avium subsp. Paratuberculosis (MAP) shedding severity in cattle based on young stock heifer faecal microbiota composition using random forest algorithms

2021 ◽  
Vol 3 (1) ◽  
Author(s):  
Alexander Umanets ◽  
Annemieke Dinkla ◽  
Stephanie Vastenhouw ◽  
Lars Ravesloot ◽  
Ad P. Koets
Keyword(s):  
Random Forest ◽  
Mycobacterium Avium ◽  
Clinical Symptoms ◽  
Classification Model ◽  
Economic Losses ◽  
Diagnostic Tools ◽  
Microbiota Composition ◽  
Faecal Microbiota ◽  
Microbial Composition ◽  
Model Based

Abstract Background Bovine paratuberculosis is a devastating infectious disease caused by Mycobacterium avium subsp. paratuberculosis (MAP). The development of the paratuberculosis in cattle can take up to a few years and vastly differs between individuals in severity of the clinical symptoms and shedding of the pathogen. Timely identification of high shedding animals is essential for paratuberculosis control and minimization of economic losses. Widely used methods for detection and quantification of MAP, such as culturing and PCR based techniques rely on direct presence of the pathogen in a sample and have little to no predictive value concerning the disease development. In the current study, we investigated the possibility of predicting MAP shedding severity in cattle based on the faecal microbiota composition. Twenty calves were experimentally infected with MAP and faecal samples were collected biweekly up to four years of age. All collected samples were subjected to culturing on selective media to obtain data about shedding severity. Faecal microbiota was profiled in a subset of samples (n = 264). Using faecal microbiota composition and shedding intensity data a random forest classifier was built for prediction of the shedding status of the individual animals. Results The results indicate that machine learning approaches applied to microbial composition can be used to classify cows into groups by severity of MAP shedding. The classification accuracy correlates with the age of the animals and use of samples from older individuals resulted in a higher classification precision. The classification model based on samples from the first 12 months of life showed an AUC between 0.78 and 0.79 (95% CI), while the model based on samples from animals older than 24 months showed an AUC between 0.91 and 0.92 (95% CI). Prediction for samples from animals between 12 and 24 month of age showed intermediate accuracy [AUC between 0.86 and 0.87 (95% CI)]. In addition, the results indicate that a limited number of microbial taxa were important for classification and could be considered as biomarkers. Conclusions The study provides evidence for the link between microbiota composition and severity of MAP infection and shedding, as well as lays ground for the development of predictive diagnostic tools based on the faecal microbiota composition.

scholarly journals Transformer Oil Quality Assessment Using Random Forest with Feature Engineering

Energies ◽  
2021 ◽  
Vol 14 (7) ◽  
pp. 1809
Author(s):  
Mohammed El Amine Senoussaoui ◽  
Mostefa Brahami ◽  
Issouf Fofana
Keyword(s):  
Machine Learning ◽  
Random Forest ◽  
Oil Quality ◽  
Principal Component ◽  
Condition Assessment ◽  
Classification Performance ◽  
Transformer Oil ◽  
Classification Model ◽  
Insulation Degradation ◽  
Transformer Oils

Machine learning is widely used as a panacea in many engineering applications including the condition assessment of power transformers. Most statistics attribute the main cause of transformer failure to insulation degradation. Thus, a new, simple, and effective machine-learning approach was proposed to monitor the condition of transformer oils based on some aging indicators. The proposed approach was used to compare the performance of two machine-learning classifiers: J48 decision tree and random forest. The service-aged transformer oils were classified into four groups: the oils that can be maintained in service, the oils that should be reconditioned or filtered, the oils that should be reclaimed, and the oils that must be discarded. From the two algorithms, random forest exhibited a better performance and high accuracy with only a small amount of data. Good performance was achieved through not only the application of the proposed algorithm but also the approach of data preprocessing. Before feeding the classification model, the available data were transformed using the simple k-means method. Subsequently, the obtained data were filtered through correlation-based feature selection (CFsSubset). The resulting features were again retransformed by conducting the principal component analysis and were passed through the CFsSubset filter. The transformation and filtration of the data improved the classification performance of the adopted algorithms, especially random forest. Another advantage of the proposed method is the decrease in the number of the datasets required for the condition assessment of transformer oils, which is valuable for transformer condition monitoring.

scholarly journals Extraction of Arecanut Planting Distribution Based on the Feature Space Optimization of PlanetScope Imagery

Agriculture ◽  
2021 ◽  
Vol 11 (4) ◽  
pp. 371
Author(s):  
Yu Jin ◽  
Jiawei Guo ◽  
Huichun Ye ◽  
Jinling Zhao ◽  
Wenjiang Huang ◽  
...  
Keyword(s):  
Random Forest ◽  
Satellite Imagery ◽  
Feature Space ◽  
Kappa Coefficient ◽  
Classification Model ◽  
Support Vector ◽  
Textural Feature ◽  
Monitoring Accuracy ◽  
Areca Catechu ◽  
High Level

The remote sensing extraction of large areas of arecanut (Areca catechu L.) planting plays an important role in investigating the distribution of arecanut planting area and the subsequent adjustment and optimization of regional planting structures. Satellite imagery has previously been used to investigate and monitor the agricultural and forestry vegetation in Hainan. However, the monitoring accuracy is affected by the cloudy and rainy climate of this region, as well as the high level of land fragmentation. In this paper, we used PlanetScope imagery at a 3 m spatial resolution over the Hainan arecanut planting area to investigate the high-precision extraction of the arecanut planting distribution based on feature space optimization. First, spectral and textural feature variables were selected to form the initial feature space, followed by the implementation of the random forest algorithm to optimize the feature space. Arecanut planting area extraction models based on the support vector machine (SVM), BP neural network (BPNN), and random forest (RF) classification algorithms were then constructed. The overall classification accuracies of the SVM, BPNN, and RF models optimized by the RF features were determined as 74.82%, 83.67%, and 88.30%, with Kappa coefficients of 0.680, 0.795, and 0.853, respectively. The RF model with optimized features exhibited the highest overall classification accuracy and kappa coefficient. The overall accuracy of the SVM, BPNN, and RF models following feature optimization was improved by 3.90%, 7.77%, and 7.45%, respectively, compared with the corresponding unoptimized classification model. The kappa coefficient also improved. The results demonstrate the ability of PlanetScope satellite imagery to extract the planting distribution of arecanut. Furthermore, the RF is proven to effectively optimize the initial feature space, composed of spectral and textural feature variables, further improving the extraction accuracy of the arecanut planting distribution. This work can act as a theoretical and technical reference for the agricultural and forestry industries.

Sign in / Sign up

Export Citation Format

Share Document