PWA-PEM for Latent Tree Model and Hierarchical Topic Detection

We propose a novel method for approximate inference in Bayesian networks (BNs). The idea is to sample data from a BN, learn a latent tree model (LTM) from the data offline, and when online, make inference with the LTM instead of the original BN. Because LTMs are tree-structured, inference takes linear time. In the meantime, they can represent complex relationship among leaf nodes and hence the approximation accuracy is often good. Empirical evidence shows that our method can achieve good approximation accuracy at low online computational cost.

Download Full-text

A Survey on Latent Tree Models and Applications

Journal of Artificial Intelligence Research ◽

10.1613/jair.3879 ◽

2013 ◽

Vol 47 ◽

pp. 157-203 ◽

Cited By ~ 24

Author(s):

R. Mourad ◽

C. Sinoquet ◽

N. L. Zhang ◽

T. Liu ◽

P. Leray

Keyword(s):

Graphical Models ◽

Latent Variables ◽

Simple Structure ◽

Tree Model ◽

The Past ◽

Latent Tree Models ◽

Tree Models ◽

Complex Relationships ◽

Latent Tree Model ◽

Comprehensive Study

In data analysis, latent variables play a central role because they help provide powerful insights into a wide variety of phenomena, ranging from biological to human sciences. The latent tree model, a particular type of probabilistic graphical models, deserves attention. Its simple structure - a tree - allows simple and efficient inference, while its latent variables capture complex relationships. In the past decade, the latent tree model has been subject to significant theoretical and methodological developments. In this review, we propose a comprehensive study of this model. First we summarize key ideas underlying the model. Second we explain how it can be efficiently learned from data. Third we illustrate its use within three types of applications: latent structure discovery, multidimensional clustering, and probabilistic inference. Finally, we conclude and give promising directions for future researches in this field.

Download Full-text

Verifying TCM Syndrome Hypothesis Based on Improved Latent Tree Model

Intelligent Computing Theories and Application - Lecture Notes in Computer Science ◽

10.1007/978-3-319-95933-7_55 ◽

2018 ◽

pp. 460-469

Author(s):

Nian Zhou ◽

Lingshan Zhou ◽

Lili Peng ◽

Bing Wang ◽

Peng Chen ◽

...

Keyword(s):

Tree Model ◽

Tcm Syndrome ◽

Latent Tree Model

Download Full-text

Why does the Dyad-4PNO model of Kern and Culpepper (2020) fit real data?

10.31234/osf.io/razq9 ◽

2020 ◽

Author(s):

Gunter Maris ◽

Timo Bechger ◽

Benjamin Deonovic

Keyword(s):

Simple Proof ◽

Real Data ◽

Psychological Measurement ◽

Tree Model ◽

Latent Tree Model

This note aims to elucidate why the Dyad-4PNO model of Kern and Culpepper (2020) can be expected to fit real data reasonably well. The main result is that the Dyad-4PNO approximates a latent tree model. We offer a simple proof of identifiability, and draw some implications for psychological measurement in practice.

Download Full-text

Numerical simulation of the tree higro-thermal response in forest fire environment

WEENTECH Proceedings in Energy ◽

10.32438/wpe.0220 ◽

2020 ◽

pp. 57-65

Author(s):

Eusébio Conceiçã ◽

João Gomes ◽

Maria Manuela Lúcio ◽

Jorge Raposo ◽

Domingos Xavier Viegas ◽

...

Keyword(s):

Forest Fire ◽

Numerical Study ◽

Thermal Response ◽

View Factor ◽

Tree Model ◽

Pine Tree ◽

Surface Temperatures ◽

Fire Environment ◽

Fire Front ◽

Heat Exchanges

This paper refers to a numerical study of the hypo-thermal behaviour of a pine tree in a forest fire environment. The pine tree thermal response numerical model is based on energy balance integral equations for the tree elements and mass balance integral equation for the water in the tree. The simulation performed considers the heat conduction through the tree elements, heat exchanges by convection between the external tree surfaces and the environment, heat exchanges by radiation between the flame and the external tree surfaces and water heat loss by evaporation from the tree to the environment. The virtual three-dimensional tree model has a height of 7.5 m and is constituted by 8863 cylindrical elements representative of its trunks, branches and leaves. The fire front has 10 m long and a 2 m high. The study was conducted taking into account that the pine tree is located 5, 10 or 15 m from the fire front. For these three analyzed distances, the numerical results obtained regarding to the distribution of the view factors, mean radiant temperature and surface temperatures of the pine tree are presented. As main conclusion, it can be stated that the values of the view factor, MRT and surface temperatures of the pine tree decrease with increasing distance from the pine tree in front of fire.

Download Full-text

The Arabic Origins of English and Indo-European “Definite Articles”: A Radical Linguistic Theory Approach

SMART MOVES JOURNAL IJELLH ◽

10.24113/ijellh.v4i6.1489 ◽

2016 ◽

Vol 4 (6) ◽

pp. 26

Author(s):

Zaidan Ali Jassem

Keyword(s):

Comparative Method ◽

Linguistic Theory ◽

General Level ◽

Theory Approach ◽

Tree Model ◽

Linguistic Change ◽

European Languages ◽

The Family ◽

Different Types ◽

Definite Articles

This paper traces the Arabic origins or cognates of the “definite articles” in English and Indo-European languages from a radical linguistic (or lexical root) theory perspective. The data comprises the definite articles in English, German, French, Spanish, Portuguese, Italian, Romanian, Latin, Greek, Macedonian, Russian, Polish, Sanskrit, Hindi, Bengali, Persian, and Arabic. The results clearly indicate that five different types of such articles emerged in the data, all of which have true Arabic cognates with the same or similar forms and meanings, whose differences are due to natural and plausible causes and different routes of linguistic change, especially lexical, semantic, or morphological shift. Therefore, the results support the adequacy of the radical linguistic theory according to which, unlike the Family Tree Model or Comparative Method, Arabic, English, German, French, Latin, Greek, and Sanskrit not only belong to the same language family, renamed Eurabian or Urban family, but also are dialects of the same language, with Arabic being their origin all because only it shares the whole cognates with them all and because it has a huge phonetic, morphological, grammatical, and lexical variety. They also manifest fundamental flaws and grave drawbacks which plague English and Indo-European lexicography for ignoring Arabic as an ultimate ancestor and progenitor not only in the treatment of the topic at hand but in all others in general. On a more general level, they also show that there is a radical language from which all human languages stemmed and which has been preserved almost intact in Arabic, thus being the most conservative and productive language

Download Full-text

Assessing the Risk of Noncompliance in Wastewater Treatment

Water Science & Technology ◽

10.2166/wst.1992.0584 ◽

1992 ◽

Vol 26 (5-6) ◽

pp. 1411-1420 ◽

Cited By ~ 2

Author(s):

S. H. Choudhury ◽

S. L. Yu ◽

Y. Y. Haimes

Keyword(s):

Wastewater Treatment ◽

Wastewater Treatment Plant ◽

Fault Tree ◽

Treatment Plant ◽

Risk Identification ◽

System Component ◽

System Failure ◽

Tree Model ◽

Risk Quantification ◽

Failure Probabilities

This paper presents an integrated methodology that allows determining the probability of noncompliance for a given wastewater treatment plant. The methodology applies fault-tree analysis, which uses failure probabilities of individual components, to predict the overall system failure probability. The methodology can be divided into two parts : risk identification and risk quantification. In risk identification, the key components in the system are determined by analyzing the contribution of individual component failures toward system failure (i.e., noncompliance). In risk quantification, a fault-tree model is constructed for the particular system, component failure probabilities are estimated, and the fault-tree model is evaluated to determine the probability of occurrence of the top event (i.e., noncompliance). A list can be developed that ranks critical events on the basis of their contributions to the probability of noncompliance. Such a ranking should assist managers to determine which components require most attention for a better performance of the entire system. A wastewater treatment plant for treating metal-bearing rinse water from an electroplating industry is used as an example to demonstrate the application of this methodology.

Download Full-text