Time Series Forecasting with Missing Values

Wu, Su‐Mei; Chang, Chia-Yung; Lee, Shie-Jue

doi:10.4108/icst.iniscom.2015.258269

Cited by 27 publications

(11 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On top of that, during night times, PV solar systems are turned off thus only data points during sunlight hours are presented in the database. Along with the fact that we don't have enough close by PV solar output to be used for interpolating value through spatial correlation, hence in order to construct valid training dataset where the training target solar invertor output level are highly fragmented in time, we applied a technique called Local Time Index (LTI) [15], where the dataset are treated as segments of discrete events in time with time index as variables joint together to form event records, replacing a fixed interval continue stream of time series data. By matching these target time label with weather condition data, time depended training dataset with event records can be generated for the machine learning modules to train.…”

Section: Preprocessingmentioning

confidence: 99%

Solar Power Prediction via Support Vector Machine and Random Forest

Yen

Hsieh

et al. 2018

E3S Web Conf.

View full text Add to dashboard Cite

Due to the variability and instability of photovoltaic (PV) output, the accurate prediction of PV output power plays a major role in energy market for PV operators to optimize their profits in energy market. In order to predict PV output, environmental parameters such as temperature, humidity, rainfall and win speed are gathered as indicators and different machine learning models are built for each solar panel inverters. In this paper, we propose two different kinds of solar prediction schemes for one-hour ahead forecasting of solar output using Support Vector Machine (SVM) and Random Forest (RF).

show abstract

Section: Preprocessingmentioning

confidence: 99%

Solar Power Prediction via Support Vector Machine and Random Forest

Yen

Hsieh

et al. 2018

E3S Web Conf.

View full text Add to dashboard Cite

show abstract

“…A lot of models of data mining/statistics have been proposed to reconstruct the missing data in a single time series thereby provide a dependable big data analytics [33,34]. SVM models [16] proposed by Frasconi et al adopt a "seasonal kernel" to estimate the similarity between different time-series.…”

Section: Missing Data Reconstruction For Cps Time Series Analyticsmentioning

confidence: 99%

A Dependable Time Series Analytic Framework for Cyber-Physical Systems of IoT-based Smart Grid

Wang

Zhu

Shi

et al. 2018

ACM Trans. Cyber-Phys. Syst.

View full text Add to dashboard Cite

With the emergence of cyber-physical systems (CPS), we are now at the brink of next computing revolution. IoT (Internet of Things) based Smart Grid (SG) is one of the foundations of this CPS revolution and defined as a power grid integrated with a large network of smart objects. The volume of time series of SG equipment is tremendous and the raw time series are very likely to contain missing values because of undependable network transferring. The problem of storing a tremendous volume of raw time series thereby providing a solid support for precise time series analytics now becomes tricky. In this paper, we propose a dependable time series analytics (DTSA) framework for IoT-based SG. Our proposed DTSA framework is capable of proving a dependable data transforming from CPS to the target database with an extraction engine to preliminary refining raw data and further cleansing the data with a correction engine built on top of a sensor-networkregularization based matrix factorization (SnrMF) method. The experimental results reveal that our proposed DTSA framework is capable of effectively increasing the dependability of raw time series transforming between CPS and the target database system through the online lightweight extraction engine and the offline correction engine. Our proposed DTSA framework would be useful for other industrial big data practices.

show abstract

“…Multiple correlated sensor readings were taken into account for doing imputation which increased computational efficiency and imputation accuracy.Shin-Fu Wu et al [9] proposed a new prediction method based on least squares support vector machine (LSSVM). Time series data as well as local time indexes were sent to LSSVM for performing prediction without imputation.…”

Section: Related Workmentioning

confidence: 99%

Imputing the Missing Values in Iot Using Estcp Model

Mary¹

2017

ijarcs

View full text Add to dashboard Cite

:In Internet of Things, occurrence of missing data is inevitable due to its intrinsic characteristics. This missing data phenomenon occurs due to a variety of reasons such as uneven network communication, synchronization difficulties, untrustworthy sensor devices, environmental aspects and other device malfunctions which often resulted in data incompleteness. A robust approach to missing data is an indispensible component of analysis to promote the perfect explanation of research findings.As the data generated by the IoT devices is usually correlated in space and time, in this paper it is demonstrated experimentally that substituting missing sensor values with spatially and temporally correlated sensor readings using thenovel extended spatial and temporal correlated proximate missing data imputation model (ESTCP)has considerably improved the accuracy than that of the previously proposed STCP model and the existing single imputation and multiple imputation techniques.

show abstract

Time Series Forecasting with Missing Values

Cited by 27 publications

References 17 publications

Solar Power Prediction via Support Vector Machine and Random Forest

Solar Power Prediction via Support Vector Machine and Random Forest

A Dependable Time Series Analytic Framework for Cyber-Physical Systems of IoT-based Smart Grid

Imputing the Missing Values in Iot Using Estcp Model

Contact Info

Product

Resources

About