The dependency matrix is widely used in fault diagnosis as a valid diagnostic model. However, the conventional dependency matrix-based diagnosis method pays more attention to the extension and improvement of the dependency matrix, and the timeliness is rarely discussed. To solve the problem and improve the diagnosis efficiency, fast diagnosis method using two search algorithms is proposed. The hash function is introduced, and the dependency matrix is transformed into the key-value pairs. The binary search algorithm is brought in to locate the desired failure mode by splitting the dependency matrix. Thus, the fault diagnosis is converted into a search process, which can improve the diagnosis efficiency by making full use of the advantages of the search algorithms. Finally, a numerical simulation is utilized to verify the proposed method. The proposed methods are applied to matrices of different sizes. The results show that the hash-based diagnosis method has significantly better efficiency, making it more suitable for onboard diagnosis. On the other hand, the binary search-based method is advantageous from a memory perspective only when the matrix size is small. 

Dependency Matrix-Based Fast Fault Diagnosis Using Search Algorithm

The dependency matrix is widely used in fault diagnosis as a valid diagnostic model. However, the conventional dependency matrix-based diagnosis method pays more attention to the extension and improvement of the dependency matrix, and the timeliness is rarely discussed. To solve the problem and improve the diagnosis efficiency, fast diagnosis method using two search algorithms is proposed. The hash function is introduced, and the dependency matrix is transformed into the key-value pairs. The binary search algorithm is brought in to locate the desired failure mode by splitting the dependency matrix. Thus, the fault diagnosis is converted into a search process, which can improve the diagnosis efficiency by making full use of the advantages of the search algorithms. Finally, a numerical simulation is utilized to verify the proposed method. The proposed methods are applied to the matrices of different sizes. The results show that the hash-based diagnosis method has significantly better efficiency, making it more suitable for onboard diagnosis. On the other hand, the binary search-based method is advantageous from a memory perspective only when the matrix size is small.

As a systematic diagnosis model, the dependency matrix describes the relationship between failure mode and test, and is widely utilized in the industry. In engineering, the dependency matrix is transformed or reshaped for different usages, and the related dependency matrix-based methods are various and complicated. However, the related approaches are not summarized systematically. Here, we provide a concise and comprehensive review of the related approaches, which are concluded into three categories: dependency matrix generation, test selection, and fault diagnosis. For each part, the conventional and the latest methods are introduced. This review is dedicated to introducing the current status of the dependency matrix-based method in fault diagnosis. We hope the review can provide an effective direction to overcome the challenges that the current methods face.

A review on dependency matrix and its application in fault diagnosis

As a systematic diagnosis model, the dependency matrix describes the relationship between failure mode and test, and is widely utilized in the industry. In engineering, the dependency matrix is transformed or reshaped for different usages, and the related dependency matrix-based methods are various and complicated. However, the related approaches are not summarized systematically. Here, we provide a concise and comprehensive review of the related approaches, which are concluded into three categories: dependency matrix generation, test selection, and fault diagnosis. For each part, the conventional and the latest methods are introduced. This review is dedicated to introducing the current status of the dependency matrix-based method in fault diagnosis. We hope the review can provide an effective direction to overcome the challenges that the current methods face. 

In recent years, data mining and text mining techniques have been frequently used for analyzing data. Electronic data is collected in everywhere and many products and services are widely used in our daily lives. Data mining techniques such as association analysis and cluster analysis are used for marketing analysis, because those can discover relationships and rules hiding in enormous numerical data. On the other hand, text mining techniques such as keywords extraction and opinion extraction are used for questionnaire or review text analysis, because those can support us to investigate consumers’ opinion in text data. However, data mining tools and text mining tools cannot be used in a single environment. Therefore, a data which has both numerical and text data is not well analyzed because the numerical part and the text part cannot be connected for interpretation. Goal of the data analysis is knowledge emergence that we find or create a new knowledge for decision making. In this paper, a mining framework that can treat both numerical and text data is proposed. Users of the proposed system can iterate data shrink and data analysis with both numerical and text analysis tools in a unique framework. Based on the experimental results, the proposed system was effectively used to data analysis for review texts of humidifiers and fan heaters. We verified that balanced use of numerical and text analysis leads to good ideas and the users should be conscious to use both type of tools and both type of data shrink.

/pdf/data-mining-framework-for-treating-both-numerical-and-text-e8j7k7ke7v.pdf

Data Mining Framework for Treating both Numerical and Text Data

IEEE Standard 1232 provides the D-matrix for diagnosing quality in models. The framework give the ability to detect dependency in relation to symptoms and failure modes [1]. This paper describes an approach to construct D-matrix by mining unstructured repair verbatim text. At first d-matrix is constructed for different dataset, and then we can form a combined d-matrix from different dataset to identify common patterns in it. In this proposed method training is performed by using different classification methods on unstructured verbatim (Combined D-Matrix) collected from the medical domain.

Text Mining Methodology to Build Dependency Matrix from Unstructured Text to Perform Fault Diagnosis

As we know that there are many doctors in all around the world, each doctor treats their patients based on their knowledge and skills. So there may be a possibility of misdiagnosis or sometimes rare diseases may be unnoticed by the doctors. It is really very hectic process for most of the doctors as they have to provide ICD-10 CM code after diagnosis. Hence, we have proposed a predictive system to identify diseases based on symptoms and give best treatment to the patient using text mining and machine learning algorithms. Our system identifies the generalized diseases and provides treatment. Our system is helpful for those patients who cannot afford to go to the physician for treatment. Also, our system is useful for doctors as it makes their work easy as we will provide ICD-10 CM code after classifying the disease. Here, in our system, the patient provides unstructured data as an input to the system and the system will determine diseases from the data provided by the system and suggest the best suitable drugs to treat their diseases. Our models consist of classifiers such as Random Forest, Naive Bayes, Support Vector Machine and Logistic Regression machine learning algorithms. Based on the performance of the classifiers we are doing the comparative analysis to know which classifier is more accurate based on their accuracy score. 

Ashish Ramdasi

Papers

Text Mining Methodology to Build Dependency Matrix from Unstructured Text to Perform Fault Diagnosis

Text Mining and Machine Learning Algorithms to Identifying Diseases and Providing Repair Action Using ICD-10 Codes