A New Neural Network to Process Missing Data without Imputation

Randolph-Gips, Mary M.

doi:10.1109/icmla.2008.89

Cited by 4 publications

(3 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this case, good performance is achieved when the collected data are sufficiently large and the classes are evenly distributed. However, there are sometimes missing values in the sensor data collected for equipment failure, maintenance, and the repair of equipment [1], and missing values in the data obtained in real time affect the final performance of machine learning models [2]. In addition, during the manufacturing process, a data imbalance can occur when there are many samples of good products and insufficient data for samples of defective products.…”

Section: Introductionmentioning

confidence: 99%

Data Preprocessing Combination to Improve the Performance of Quality Classification in the Manufacturing Process

2022

View full text Add to dashboard Cite

The recent introduction of smart manufacturing, also called the ‘smart factory’, has made it possible to collect a significant number of multi-variate data from Internet of Things devices or sensors. Quality control using these data in the manufacturing process can play a major role in preventing unexpected time and economic losses. However, the extraction of information about the manufacturing process is limited when there are missing values in the data and a data imbalance set. In this study, we improve the quality classification performance by solving the problem of missing values and data imbalances that can occur in the manufacturing process. This study proceeds with data cleansing, data substitution, data scaling, a data balancing model methodology, and evaluation. Five data balancing methods and a Generative Adversarial Network (GAN) were used to proceed with data imbalance processing. The proposed schemes achieved an F1 score that was 0.5 higher than the F1 score of previous studies that used the same data. The data preprocessing combination proposed in this study is intended to be used to solve the problem of missing values and imbalances that occur in the manufacturing process.

show abstract

Section: Introductionmentioning

confidence: 99%

Data Preprocessing Combination to Improve the Performance of Quality Classification in the Manufacturing Process

2022

View full text Add to dashboard Cite

show abstract

“…Therefore, it is not immediately obvious how one would use such a model in the second category of techniques (i.e., without interpolation of the gaps) and this remains an open problem in the machine learning community (Caiafa et al., 2021; Emmanuel et al., 2021; Sharpe & Solly, 1995). One solution is to use a Cosine Neural Network (Randolph‐Gips, 2008). This architecture is able to process and recognize missing data without any prior imputation, thereby addressing the issue of how to represent missing data to a neural network.…”

Section: Introductionmentioning

confidence: 99%

Exploring the Potential of Neural Networks to Predict Statistics of Solar Wind Turbulence

et al. 2022

View full text Add to dashboard Cite

Time series data sets often have missing or corrupted entries, which need to be handled in subsequent data analysis. For example, in the context of space physics, calibration issues, satellite telemetry issues, and unexpected events can make parts of a time series unusable. This causes problems for understanding the dynamics of the heliosphere and space weather environment. Various approaches exist to tackle this problem, including mean/median imputation, linear interpolation, and autoregressive modeling. Here, we study the utility of artificial neural networks (ANNs) to predict statistics of sparse time series. Our focus is not on time series prediction but on gleaning the best possible information about the statistical behavior of the system. As an example application, we focus on the structure functions of turbulent time series measured in the solar wind. Using a data set with artificial gaps, a neural network is trained to predict second‐order structure functions and then tested on an unseen data set to quantify its performance. A small feedforward ANN, with only 20 hidden neurons, can predict the large‐scale fluctuation amplitudes better than mean imputation or linear interpolation when the percentage of missing data is high. Although they perform worse than the other methods when it comes to capturing both the shape and fluctuation amplitude together, their performance is better in a statistical sense for large fractions of missing data. Caveats regarding their utility, the optimization procedure, and potential future improvements are discussed.

show abstract

“…When utilizing clinical data, physicians will not perform all tests on all patients, resulting in missing data when patients are combined. Therefore techniques need to be utilized to make the most of the data that is present [165,166]. …”

Section: Introductionmentioning

confidence: 99%

Modeling autism: a systems biology approach

Randolph-Gips

Srinivasan²

2012

J Clin Bioinformatics

View full text Add to dashboard Cite

Autism is the fastest growing developmental disorder in the world today. The prevalence of autism in the US has risen from 1 in 2500 in 1970 to 1 in 88 children today. People with autism present with repetitive movements and with social and communication impairments. These impairments can range from mild to profound. The estimated total lifetime societal cost of caring for one individual with autism is $3.2 million US dollars. With the rapid growth in this disorder and the great expense of caring for those with autism, it is imperative for both individuals and society that techniques be developed to model and understand autism. There is increasing evidence that those individuals diagnosed with autism present with highly diverse set of abnormalities affecting multiple systems of the body. To this date, little to no work has been done using a whole body systems biology approach to model the characteristics of this disorder. Identification and modelling of these systems might lead to new and improved treatment protocols, better diagnosis and treatment of the affected systems, which might lead to improved quality of life by themselves, and, in addition, might also help the core symptoms of autism due to the potential interconnections between the brain and nervous system with all these other systems being modeled. This paper first reviews research which shows that autism impacts many systems in the body, including the metabolic, mitochondrial, immunological, gastrointestinal and the neurological. These systems interact in complex and highly interdependent ways. Many of these disturbances have effects in most of the systems of the body. In particular, clinical evidence exists for increased oxidative stress, inflammation, and immune and mitochondrial dysfunction which can affect almost every cell in the body. Three promising research areas are discussed, hierarchical, subgroup analysis and modeling over time. This paper reviews some of the systems disturbed in autism and suggests several systems biology research areas. Autism poses a rich test bed for systems biology modeling techniques.

show abstract

A New Neural Network to Process Missing Data without Imputation

Cited by 4 publications

References 7 publications

Data Preprocessing Combination to Improve the Performance of Quality Classification in the Manufacturing Process

Data Preprocessing Combination to Improve the Performance of Quality Classification in the Manufacturing Process

Exploring the Potential of Neural Networks to Predict Statistics of Solar Wind Turbulence

Modeling autism: a systems biology approach

Contact Info

Product

Resources

About