1. bookAHEAD OF PRINT
Detalles de la revista
License
Formato
Revista
eISSN
2444-8656
Primera edición
01 Jan 2016
Calendario de la edición
2 veces al año
Idiomas
Inglés
Acceso abierto

English Intelligent Question Answering System Based on elliptic fitting equation

Publicado en línea: 15 Jul 2022
Volumen & Edición: AHEAD OF PRINT
Páginas: -
Recibido: 28 Apr 2022
Aceptado: 22 Jun 2022
Detalles de la revista
License
Formato
Revista
eISSN
2444-8656
Primera edición
01 Jan 2016
Calendario de la edición
2 veces al año
Idiomas
Inglés
Introduction

With the rapid development of network communication and the arrival of big data, it has entered the human brain step by step. The Internet is flooded with hard data, and the traffic of information on the Internet is constantly increasing. This vast amount of information has become the key to answering consumer questions. Most SEO searches by keyword. While this retrieval method makes it easier for users to store data, it returns some search results and does not affect users. It is difficult for users to quickly find the answers they are looking for. How to obtain accurate data to meet most of the big data needs of consumers is the goal of scientists in the information age. In recent years, the development of question answering systems has gradually attracted the attention of researchers. Intelligent question answering system is a new research in natural language processing and data recovery. It allows users to ask questions in natural sentences and uses natural language tools to get short and accurate answers for users. It was created by humans to find information quickly and accurately. In addition to comparing user queries with existing search queries, smart queries have the best results in understanding queries compared to search engine contextual comparison methods and can directly improve answers. Users want them to be presented to the user in a certain way. Answering these smart answers plays an important role in answering distance learning questions and solving challenges [1].

Data ontology is the basis for users to “answer questions”, and appropriate answers to these questions are stored in the database. When a user asks a question, the question answering system usually deletes the common question and returns it to the user to see similar answers. If the system cannot find enough answers from the FAQ database, the system will switch to the ontology database. Among them, algorithms are techniques for answering key questions for finding similar questions by understanding the ontology database. After counting the similarity between the user's question and the question answer stored in the ontology information, select the answer similar to the user's question and submit the answer similar to the user [2].

Phrase similarity algorithms are now widely used in practice. Research status affects research progress in other relationships. Sentence algorithms play an important role in many query answering systems. For example, my country's query answering, ontology question answering system, OTC question answering system, multi-question-based machine question answering system, sentence-like algorithm, etc. are all important technologies in answering questions. Based on this, this document uses the similar sentence algorithm in the original English intelligent question answering system to design and implement the English intelligent question answering system. The sentence-like program completely solves the vocabulary and content of English questions, not only improving the performance of English intelligent answering, but also improving the level of students. Start answering English questions correctly. The system meets the requirements of primary school students' question and answer, and helps primary school students learn English better [3].

K. Boten et al. believe that Key Answer Questions (also known as Intelligent Question Answering Robots) is a new retrieval system that allows users to ask questions in any language without sharing the questions as keywords. The system integrates tools that work in a good language and preserves similar answers by understanding the question [4]. Shi M et al. found that the key questions can be divided into knowledge-based question answering, query-based database query, query-based query, question answering, knowledge-based answering and answering. System based Ontology [5] Nandy S et al found that the key question answering system generally includes three parts: question analysis, data recovery and retrieval. There are several practical answering techniques in restricting the answering of key questions, such as FAQ databases, writing ontology databases and corpus queries, and semantic understanding methods based on multiple equivalence metrics [6].

Method
System design

The English Language Proficiency Test is a great tool for finding information in a natural environment. The goal is to help elementary school students ask questions about simple English sentences in their daily life and then get the correct answers. In short, the intelligent English question answering system examines the English questions raised by students, understands the meaning of the questions, and then sends the answers to the students, so that students can learn to learn. As shown in Figure 1, the English intelligent question answering system has four modules: one is the question analysis module, which has only five parts: answer words, voice annotation, question sample analysis, calculation, question and answer analysis, and keywords. analyze. depending on. The second module is a similar counting module and is the core of the query. This chapter usually includes counting similar words and counting similar sentences. The third module decomposes answers and is often used to partition consistency, filter answers, and create answers. The ultimate module is to gain experience. This often provides answers to users in multiple ways by understanding their different experience levels, stressors, and user choices.

Figure 1

English Intelligent Question Answering System Model

The English Intelligence Questionnaire is data on the use of technology in natural environments. Its goal is to hope that elementary school students can ask questions about simple English sentences in their daily life and then get the correct answers. In short, the English intelligent answering system analyzes the English questions given by the students, understands the meaning of the questions, and then sends the answers to the students. As shown in Figure 1, the English intelligent answering system consists of four modules: one is the query analysis module, which usually has five parts: word correction, partial voice annotation, query type analysis, question-and-answer type analysis and keywords. Delete. The second module is a similar calculation and is essentially a quiz. This is usually similar Chinese and similar sentences. The third module, field extraction, is only used to identify connections, filter answers, and publish answers. The ultimate module is advancing knowledge. This usually provides answers for different types of users by understanding the level of knowledge of different users, the difficulty of previous experiences, and users' preferences for different types.

Error correction modules are usually based on machine language models and similar measurement models. The technical model accepts the standard model. The standard measurement model usually uses a query vector generation model to measure the query statement input by the user through intelligent error correction. It will correct the error if it matches the default. The grammar used in intelligent error correction is divided into pinyin grammar and grammar. The main function of the Pinyin language model is to correct inaccuracies, inaccuracies, etc. in the query sentences entered by the user, and the main function of the script is to correct incorrect sentences and other problems in the input sentences. By user [7].

The question analysis module is an integral part of the intelligent question answering system. The purpose is for the computer to understand the semantics of the user query and prepare the next answer for the decryption module. Accurate question analysis helps the system use methods and strategies to answer various types of questions in the answering module. When users submit English questions, the question analysis module analyzes and develops them. The development plan includes:

Word reduction changes all the words in the question and answer sentences back to the prototype. For example, the auxiliary verbs had and has become have, the plural word sports becomes sport, and the verbs be become is and was become be.

Part of speech tagging marks the part of speech of each restored word. It includes words with part of speech as verbs, words with part of speech as nouns, words with part of speech as adjectives and words with part of speech as adverbs.

Question type analysis analyzes some common English question types put forward by users, such as “What kind of...,” How do you like...?” “Which is ......?” and so on.

Q&A type review to determine what type of product a user's inventory falls under. Types of questions and answers include answers, discussion questions, etc. Each format has four media formats: audio, video, album and image. It is divided into three levels: hard, medium and easy. Also, users can write a question to benefit. Answer types include answer, sentence, sentence, suffix, etc. Distributing and translating words and questions and answers can not only improve English, but also improve students' motivation and interest in learning. Learning English is for one thing.

Keyword extraction the keywords contained in English questions represent the main meaning of the sentence. In the retrieval process, keywords are given different weights according to different parts of speech. The weights are nouns, restrictive adverbs, adjectives and verbs from large to small.

The process of calculating similar sentences is usually to select the appropriate sentence by calculating the similarity between two sentences. Sentences usually include words, syntax, and semantics. The larger the similarity value in the calculation, the closer the morphological, syntactic and semantic data of the two sentences are. In this study, a distance-based similarity algorithm was used to calculate the similarity of English questions. The algorithm is based on a semantic taxonomy dictionary of pure terms. The similarity of English words is calculated by calculating the similarity of the program. After the similarity of English words is obtained, the similarity of English meaning is obtained by statistical angular cosine similarity [8].

The unpacking module analyzes the lexical, syntactic and semantic aspects of the other answer questions obtained from the data retrieval module and needs to rank the answers. Also, the system must be set to boot. Search results are only available if the sentence similarity is greater than the initial similarity. Extract the results by filtering keywords, remove the content that does not affect the search results, and then refine the results according to the categories of keywords, propose the most suitable answer, and return it to the user. Consumers know the level. The model simply subtracts similar answers from the ontology knowledge base by counting similar questions, then analyzes the answers, ranks the above similar answers, this 0.8 based on a similarity value, and pushes another answer to the familiar push module. Problem types can be obtained from the problem analysis module.

Elliptic equation

In recent years, scientists have proposed several ellipse detection algorithms, which can be divided into three categories: one is the ellipse fitting method based on Hough transform, the other is the fitting method based on geometric features, and the third is the fitting ellipse method. according to the least squares method. In this paper, an ellipse fitting algorithm combining ellipse symmetry and bisection method is proposed, and Halcon is used as the graph algorithm library. The method utilizes the geometric symmetry of the ellipse itself to calculate the basic control of the ellipse, and then provides the bisection method to calculate the semi-axis length, minor semi-axis and direction. The angle of the ellipse is estimated by the completion. Finally, calculate the actual parameters of the ellipse[910].

To solve the elliptic equation, five parameters need to be determined. c, d are the length of the long and short half axes of the ellipse; x1, y1 are the center coordinate of the ellipse; β is the included angle between the long half axis of the ellipse and the X axis (collectively referred to as the direction angle in this paper). The equation of the ellipse can be expressed as equation (1): [(xx1)cosβ+(yy1)sinβ]2a2=1 {{{{\left[ {\left( {x - {x_1}} \right)\cos \beta + \left( {y - {y_1}} \right)\sin \beta } \right]}^2}} \over {{a^2}}} = 1

According to the mathematical model of the ellipse, the ellipse is symmetrical about the central point, that is, as a straight line at any angle, parallel lines are tangent to the ellipse, there are two tangents la, l′a (collectively referred to as tangent pairs in this paper), there are two tangent points cb and c′b (collectively referred to as tangent pairs in this paper), and the center of the circle is located at the midpoint of the connecting line of the two tangent points. The distance of the tangent point pair is recorded as dj; When dj is the largest, the distance between the tangent points is the length of the major axis of the ellipse; Similarly, when dj is the smallest, the distance between the tangent point pair is the minor axis length of the ellipse, and the included angle αj between the tangent line and the X axis is the direction angle β of the ellipse.

Firstly, the camera is calibrated to obtain internal and external parameters. The camera collects the image of elliptical hole, and carries out image correction and preprocessing. The pixel edge of ellipse is obtained by Canny operator; The sub-pixel edge of the ellipse is obtained by Gaussian fitting.

The coordinates of the ellipse center can be calculated by equations (2) and (3). x0=12ni=0n(xi+xi) {x_0} = {1 \over {2n}}\sum\limits_{i = 0}^n {\left( {{x_i} + x_i^\prime} \right)} y0=12ni=0n(yi+yi) {y_0} = {1 \over {2n}}\sum\limits_{i = 0}^n {\left( {{y_i} + y_i^\prime} \right)} Where: x1 and y1 are the estimated ellipse center coordinates; N is the logarithm of the tangent point pair.

Estimate the semi axial length of the ellipse length a iteratively by dichotomy: In the previous step, you can get the values of n di and ai, sort the n di from small to large, and take the first two as dk, ds (dk>ds). Therefore, you can lock that the long axial length of the ellipse exists in the angle area with the tangent angle range of (ak, as). take down an angle an+1 = (ak + as)/2 by dichotomy to obtain the distance corresponding to the new tangent pair and tangent point pair: ln+1, ln+1 l_{n + 1}^\prime , dn+1. Lock the tangent angle area (an+1, an+2) where the next ellipse's long axis length appears, where: an+2 = (ak + an+1)/2. In this way, the real value approximating the ellipse's long axis length is continuously updated in the iteration through the dichotomy method. Suppose that when the iteration reaches the j-th time, the absolute value of the difference between dn+j and dn+j+1 is less than the set iteration threshold ɛ. at this time, the iteration ends, and the calculation of the ellipse's long half axis length is shown in equation (4): α=(dn+j+dn+j+1)4 \alpha = {{\left( {{d_{n + j}} + {d_{n + j + 1}}} \right)} \over 4}

Similarly, calculate the length b and direction angle θ of the short semi axis of the ellipse, use the same iterative approximation method, and sort n di from small to large through the data obtained in formula (2), and the last two are recorded as dv and dp. Thus, it can be locked that the minor axis length of the ellipse exists in the tangent angle range of (av, ap), which is the same as the iteration method in 3), and it is assumed that when the iteration reaches the m-th time. The absolute value of the difference between dn+1 and dn+m+1 is less than the set iteration threshold ɛ. at this time, the iteration ends, and the calculation of the short semi axis length and direction angle of the ellipse is shown in equations (5) and (6): b=(dn+m+dn+m+1)4 b = {{\left( {{d_{n + m}} + {d_{n + m + 1}}} \right)} \over 4} θ=(an+m+an+m+1)2 \theta = {{\left( {{a_{n + m}} + {a_{n + m + 1}}} \right)} \over 2}

System experimental analysis

The system takes MySQL database and knowledge ontology database as the background database management system. The data tables used include question table, question answer table, answer table, learning preference table, common word dictionary and word net as semantic dictionary. Adopt (spring MVC + Maven + mybatis) and other architectures under the Eclipse platform.

After the user enters a question, the system first queries the FAQ question set. If the answer is found, it will be filtered and the most appropriate answer will be returned to the user. If the answer to the question is not found in the FAQ database, query the knowledge ontology database. This stage requires a series of syntactic processing, such as word restoration, part of speech tagging and other processes to form a keyword set, then calculate the similarity of the processed English sentences, and finally return the question answer with the highest similarity to the user, presenting diversified answer content according to the user's knowledge level. The answer to the question can query the corresponding ID value in the knowledge ontology database and obtain several attribute values related to the ID value. The user ontology database stores the knowledge level, learning style and other information of each student, and finally recommends appropriate learning content to learners through the knowledge push module.

(1) Knowledge base management

Knowledge base is the knowledge center of intelligent question answering system, which is composed of classification, examples, attributes, standard questions, extended questions and standard answers. Among them, classification refers to the classification information marked by personnel or customers on each question and answer pair of the knowledge base in advance. An example refers to a collection of different questions for the same answer, and the attribute is another kind of classification information. For example, the knowledge base of Liangshan Prefecture Government intelligent question and answer system has attributes such as to-do items, query items and so on. Standard questions and standard answers refer to common knowledge Q & A pairs given by customers, and extended questions refer to different questions provided by tagging personnel for the same tagging answer [11].

(2) Ontology class management

Ontology class is the description and formal expression of real-world concepts and the relationship between concepts. It is divided into entity class and method class. Taking the field of government affairs as an example, the ID card belongs to the entity class, the method class, and the handling belongs to the method class and the ontology class. Knowledge management is realized by summarizing and inheriting ontology classes.

(3) Knowledge Topology

Knowledge Topology visually displays the knowledge in the knowledge base in the form of topology according to different dimensions. Intuitively display the knowledge of different dimensions to customers, which is convenient for customers and developers to observe the system.

(4) Knowledge dimension

The knowledge dimension module classifies, organizes and counts knowledge by defining different dimensions. Knowledge combing usually divides knowledge into four dimensions. Taking the government affairs system as an example, the system is divided into four dimensions: state, county, matters to be handled and specific problems.

(5) Hotspot management

Hot issues refer to issues concerned by users and issues related to current events. Hot issues usually become issues concerned by users. Therefore, the system supports the addition, deletion and modification of hot issues, and hot issues are preferentially recommended in the interface.

Question understanding refers to transforming the intention expressed by the query sentence entered by the user into the semantic structure that can be recognized and understood by the intelligent question answering system through natural language technology. In the intelligent question answering system described in this paper, the question understanding function mainly depends on three sub modules. The preprocessing module passes the query statements entered by the user through intelligent word segmentation.

Name entity annotation and other methods are converted into keywords and word sequences. The session management module defines the questions raised by users. The intelligent question and answer system described in this paper helps users who cannot express their intentions to express their intentions through the session management module. The error correction module allows the user to input the correct expression of the sentence and return the error correction result.

Data test results

As shown in Table 1, the experimental results are evaluated by accuracy and comprehensive similarity. A total of 453 sentences and 920 words are input in this experiment, the accuracy is 81.23% and 89.23% respectively, and the comprehensive similarity value is greater than 0.8. The test results show that the algorithm can improve the accuracy of the experimental results, and the sentence comprehensive similarity is high, which improves the system efficiency.

System data test results

Input type Number of tests Number of correct results Accuracy (%) Comprehensive similarity
sentence 453 368 81.23 0.8173
word 920 815 89.23 0.8018
Word query results

When the user enters an English word in the search box, the system will present various types of answers related to the word, such as the original text sentence, situational paragraph, extended example sentence, etc. In the process of querying words, the system first queries the FAQ database, and if the answer is found, it will be directly presented to the user; If the FAQ database has no answer, query the knowledge ontology database to present the answer to the user. For example, take the input word “ruvier” as an example to query [12].

Question query results

When the user enters a question in the search box, the system will present various types of answers related to the question, such as answer, situational dialogue, important sentence patterns, after-school words, etc. In the process of query, the system first queries the FAQ database, and if the answer is found, it will be directly presented to the user; If the FAQ database has no answer, query the knowledge ontology database to present the answer to the user. While querying the ontology database, calculate the similarity of the found answers to relevant sentences, and select the question answer with the greatest similarity as the final answer to return to the user.

The system collects and arranges more than 2000 common and universal questions of many primary school students in the process of using the question and answer system. The question answers are imported in batch. The question answers are collected through teaching materials and then stored in the knowledge ontology database. Since the system has been running for three months, it has been used by primary school students in various schools for a total of 3058 times, and the statistical analysis is carried out. As shown in Table 2, where λ(0 ≤ λ ≤ 1) is the similarity threshold set by the system. Here, three values of 0.6/ 0.7 / 0.8 are taken to measure the sentence similarity.

Proportion of sentence similarity results

λ Number of sentences Sentence percentage Number of words Percentage of words
0.8 953 21.4% 3452 89.8%
0.7 2456 69.4% 231 4.2%
0.6 283 4.0% 156 2.1%

As can be seen from table 2: (1) When λ takes different values, the number of similar questions is different. Among them, when λ = 0.8, the number of similar sentences is the highest, accounting for 70.4% of the total proportion, indicating that the system has high accuracy in using the distance based similarity algorithm. (2) When λ = 0.9, the number of similar words is the highest, reaching more than 90%. It can be seen that the system has better met the needs of practical application in terms of query words or sentences [13].

Conclusion

Although the English answer can directly complete the skill, it depends on whether the user's question can be answered correctly. This paper examines English language similarities, teaches English proficiency answer models and the capabilities and functions of each model within each model, provides design and development processes, and mathematics and uses important English questions. The answer is based on algorithms. Due to previous experience and applications, the system is able to successfully and responsibly answer user questions and provide some insight into customer issues. In addition, the speed of the system's exploration and the distance of the similarity algorithm improve the system's accuracy in answering smart English queries. Further research will improve the consistency of the algorithm and strengthen the functionality of the system to see user needs more quickly. At the same time, we will continue to record the number of clicks and page views of users in the actual use of the system, and provide special training to help users obtain information needs. Follow personal training.

Figure 1

English Intelligent Question Answering System Model
English Intelligent Question Answering System Model

Proportion of sentence similarity results

λ Number of sentences Sentence percentage Number of words Percentage of words
0.8 953 21.4% 3452 89.8%
0.7 2456 69.4% 231 4.2%
0.6 283 4.0% 156 2.1%

System data test results

Input type Number of tests Number of correct results Accuracy (%) Comprehensive similarity
sentence 453 368 81.23 0.8173
word 920 815 89.23 0.8018

Duong A T, Luong V T, Nguyen T Q. Classification of Stable Solutions to a Fractional Singular Elliptic Equation with Weight[J]. Acta Applicandae Mathematicae, 2020, 170(1):579–591. DuongA T LuongV T NguyenT Q Classification of Stable Solutions to a Fractional Singular Elliptic Equation with Weight [J] Acta Applicandae Mathematicae 2020 170 1 579 591 10.1007/s10440-020-00347-5 Search in Google Scholar

Islam M T, Tasciotti E, R Righetti. Estimation of Vascular Permeability in Irregularly Shaped Cancers Using Ultrasound Poroelastography[J]. IEEE Transactions on Biomedical Engineering, 2020, 67(4):1083–1096. IslamM T TasciottiE RighettiR Estimation of Vascular Permeability in Irregularly Shaped Cancers Using Ultrasound Poroelastography [J] IEEE Transactions on Biomedical Engineering 2020 67 4 1083 1096 10.1109/TBME.2019.292913431331877 Search in Google Scholar

Haziot S V. Study of an Elliptic Partial Differential Equation Modeling the Ocean Flow in Arctic Gyres[J]. Journal of Mathematical Fluid Mechanics, 2021, 23(2):1–9. HaziotS V Study of an Elliptic Partial Differential Equation Modeling the Ocean Flow in Arctic Gyres [J] Journal of Mathematical Fluid Mechanics 2021 23 2 1 9 10.1007/s00021-021-00584-0 Search in Google Scholar

K Boateng, Yang W, Yaro D, et al. Jacobi Elliptic Function Solutions and Traveling Wave Solutions of the (2 + 1)-Dimensional Gardner-KP Equation[J]. Mathematical Methods in the Applied Sciences, 2020, 43(6):3457–3472. BoatengK YangW YaroD Jacobi Elliptic Function Solutions and Traveling Wave Solutions of the (2 + 1)-Dimensional Gardner-KP Equation [J] Mathematical Methods in the Applied Sciences 2020 43 6 3457 3472 10.1002/mma.6131 Search in Google Scholar

Shi M. Knowledge Graph Question and Answer System for Mechanical Intelligent Manufacturing Based on Deep Learning[J]. Mathematical Problems in Engineering, 2021, 2021(2):1–8. ShiM Knowledge Graph Question and Answer System for Mechanical Intelligent Manufacturing Based on Deep Learning [J] Mathematical Problems in Engineering 2021 2021 2 1 8 10.1155/2021/6627114 Search in Google Scholar

Nandy S, Adhikari M. Intelligent Health Monitoring System for Detection of Symptomatic/Asymptomatic COVID-19 Patient[J]. IEEE Sensors Journal, 2021, PP(99):1–1. NandyS AdhikariM Intelligent Health Monitoring System for Detection of Symptomatic/Asymptomatic COVID-19 Patient [J] IEEE Sensors Journal 2021 PP 99 1 1 10.1109/JSEN.2021.3096425 Search in Google Scholar

Khushhal S, Majid A, Abbas S A, et al. Question retrieval using combined queries in community question answering[J]. Journal of Intelligent Information Systems, 2020, 55(2):307–327. KhushhalS MajidA AbbasS A Question retrieval using combined queries in community question answering [J] Journal of Intelligent Information Systems 2020 55 2 307 327 10.1007/s10844-020-00612-x Search in Google Scholar

Liu C, Tang D, H Zhu, et al. A Novel Predictive Maintenance Method Based on Deep Adversarial Learning in the Intelligent Manufacturing System[J]. IEEE Access, 2021, PP(99):1–1. LiuC TangD ZhuH A Novel Predictive Maintenance Method Based on Deep Adversarial Learning in the Intelligent Manufacturing System [J] IEEE Access 2021 PP 99 1 1 10.1109/ACCESS.2021.3069256 Search in Google Scholar

Dovis F, Ruotsalainen L, Toledo-Moreo R, et al. Ellipse Fitting Algorithm Based on Geometric Symmetry [Guest Editorial][J]. IEEE Intelligent Transportation Systems Magazine, 2020, 12(3):6–9. DovisF RuotsalainenL Toledo-MoreoR Ellipse Fitting Algorithm Based on Geometric Symmetry [Guest Editorial] [J] IEEE Intelligent Transportation Systems Magazine 2020 12 3 6 9 10.1109/MITS.2020.2994923 Search in Google Scholar

Bv N, Guddeti R. Ellipse Fitting Algorithm Based on Geometric Symmetry 4.0[J]. IEEE Transactions on Industrial Informatics, 2021, PP(99):1–1. BvN GuddetiR Ellipse Fitting Algorithm Based on Geometric Symmetry 4.0 [J] IEEE Transactions on Industrial Informatics 2021 PP 99 1 1 Search in Google Scholar

Sami B S. Intelligent Energy Management for Off-Grid Renewable Hybrid System Using Multi-Agent Approach[J]. IEEE Access, 2020, PP(99):1–1. SamiB S Intelligent Energy Management for Off-Grid Renewable Hybrid System Using Multi-Agent Approach [J] IEEE Access 2020 PP 99 1 1 10.1109/ACCESS.2019.2963584 Search in Google Scholar

Aidara S, Sagna Y. BSDEs driven by two mutually independent fractional Brownian motions with stochastic Lipschitz coefficients[J]. Applied Mathematics and Nonlinear Sciences, 2019, 4(1):151–162 AidaraS SagnaY BSDEs driven by two mutually independent fractional Brownian motions with stochastic Lipschitz coefficients [J] Applied Mathematics and Nonlinear Sciences 2019 4 1 151 162 10.2478/AMNS.2019.1.00014 Search in Google Scholar

Saouli M A. Existence of solution for Mean-field Reflected Discontinuous Backward Doubly Stochastic Differential Equation[J]. Applied Mathematics and Nonlinear Sciences, 2020, 5(2):205–216. SaouliM A Existence of solution for Mean-field Reflected Discontinuous Backward Doubly Stochastic Differential Equation [J] Applied Mathematics and Nonlinear Sciences 2020 5 2 205 216 10.2478/amns.2020.2.00038 Search in Google Scholar

Artículos recomendados de Trend MD

Planifique su conferencia remota con Sciendo