Lower Mean Absolute Error Values Research Articles

Can time-dependent double hybrid density functionals accurately predict electronic excitation energies of BODIPY compounds?

The vertical excitation energies of 12 BODIPY chromophores are benchmarked via TD-DFT, using 36 functionals covering different rungs. It was found that most TD-DFT results overestimate the excitation energies, and provide mean absolute error (MAE) values larger than 0.4 eV. The dispersion-corrected, spin-component-scaled, double-hybrid (DSD) density functionals DSD-BLYP and DSD-PBEP86 are found to have the smallest MAE values of 0.119 eV and 0.124 eV, respectively. The low MAE values of the two DSD density functionals falls in the range of errors found in the wavefunction based methods that are reported to be suitable methods for such chromophores. DSD-BLYP and DSD-PBEP86 functionals also show excellent consistency (standard deviation = 0.076 eV and 0.082 eV respectively) and good predictability (linear determination coefficient R2 = 0.933 eV and 0.917 eV respectively). MAE values of the two spin-component-scaled DH functionals of vertical fluorescence calculations are also promising (MAE around 0.12 eV).

Change Point Enhanced Anomaly Detection for IoT Time Series Data

Due to the exponential growth of the Internet of Things networks and the massive amount of time series data collected from these networks, it is essential to apply efficient methods for Big Data analysis in order to extract meaningful information and statistics. Anomaly detection is an important part of time series analysis, improving the quality of further analysis, such as prediction and forecasting. Thus, detecting sudden change points with normal behavior and using them to discriminate between abnormal behavior, i.e., outliers, is a crucial step used to minimize the false positive rate and to build accurate machine learning models for prediction and forecasting. In this paper, we propose a rule-based decision system that enhances anomaly detection in multivariate time series using change point detection. Our architecture uses a pipeline that automatically manages to detect real anomalies and remove the false positives introduced by change points. We employ both traditional and deep learning unsupervised algorithms, in total, five anomaly detection and five change point detection algorithms. Additionally, we propose a new confidence metric based on the support for a time series point to be an anomaly and the support for the same point to be a change point. In our experiments, we use a large real-world dataset containing multivariate time series about water consumption collected from smart meters. As an evaluation metric, we use Mean Absolute Error (MAE). The low MAE values show that the algorithms accurately determine anomalies and change points. The experimental results strengthen our assumption that anomaly detection can be improved by determining and removing change points as well as validates the correctness of our proposed rules in real-world scenarios. Furthermore, the proposed rule-based decision support systems enable users to make informed decisions regarding the status of the water distribution network and perform effectively predictive and proactive maintenance.

Analyzing Intracranial EEG in Pharmacoresistant Epilepsy Patients Using Hidden Markov Models and Time Series Forecasting Methods

Various brain activities can be captured by electroencephalographic signals, which can then be used to detect epilepsy considering that the epileptic seizures are caused by a disturbance in the electrophysiological activity of the brain. Accordingly, epileptic seizure prediction usually requires a careful analysis of electroencephalography (EEG) records. In this study, we examined a large intracranial EEG dataset obtained from five pharmacoresistant epilepsy patients. Specifically, we first applied hidden Markov models to parse the amplitude under a probabilistic description considering the observed data as the outcome of either one of the three hidden states, namely, normal, subclinical (seizure) and clinical (seizure), which is in line with the setup proposed in previous studies. Our results indicate that there are indeed a maximum of 1.7% subclinical and about 1% clinical events, which comply with the observations from other studies. Next, we assumed EEG signals to form complex time series, and considered time-series prediction methods to forecast the future events. Such predictions are of interest to predetermine the possibility of a seizure onset and taking a preventive strategy. This task was performed within a deterministic framework by applying deep learning methods. An ensemble model was created using one-dimensional convolution net in conjunction with long short-term memory units and deep neural networks. We observe that the proposed time series prediction method is highly accurate as indicated by the low mean absolute error values and the high conformity of the predictions to the ground truth values.

Associations of Body Mass Index with Demographics, Lifestyle, Food Intake, and Mental Health among Postpartum Women: A Structural Equation Approach.

As postpartum obesity is becoming a global public health challenge, there is a need to apply postpartum obesity modeling to determine the indicators of postpartum obesity using an appropriate statistical technique. This research comprised two phases, namely: (i) development of a previously created postpartum obesity modeling; (ii) construction of a statistical comparison model and introduction of a better estimator for the research framework. The research model displayed the associations and interactions between the variables that were analyzed using the Structural Equation Modeling (SEM) method to determine the body mass index (BMI) levels related to postpartum obesity. The most significant correlations obtained were between BMI and other substantial variables in the SEM analysis. The research framework included two categories of data related to postpartum women: living in urban and rural areas in Iran. The SEM output with the Bayesian estimator was 81.1%, with variations in the postpartum women’s BMI, which is related to their demographics, lifestyle, food intake, and mental health. Meanwhile, the variation based on SEM with partial least squares estimator was equal to 70.2%, and SEM with a maximum likelihood estimator was equal to 76.8%. On the other hand, the output of the root mean square error (RMSE), mean absolute error (MSE) and mean absolute percentage error (MPE) for the Bayesian estimator is lower than the maximum likelihood and partial least square estimators. Thus, the predicted values of the SEM with Bayesian estimator are closer to the observed value compared to maximum likelihood and partial least square. In conclusion, the higher values of R-square and lower values of MPE, RMSE, and MSE will produce better goodness of fit for SEM with Bayesian estimators.

Soft-computing-centric framework for wildfire monitoring, prediction and forecasting

Wildfires are exorbitantly cataclysmic disasters that lead to the destruction of forest cover, wildlife, land resources, human assets, reduced soil fertility and global warming. Every year wildfires wreck havoc across the globe. Therefore, there is a need of an efficient and reliable system for real-time wildfire monitoring to dilute their disastrous effects. Internet of Things (IoT) has demonstrated remarkable evolution and has been successfully adopted in environmental monitoring domain. This paper proposes a collaborative IoT–Fog–Cloud framework based on soft computing techniques for real-time wildfire monitoring, prediction and forecasting. The framework includes proposals for classifying a forest terrain into its appropriate wildfire proneness class using fuzzy K-nearest-neighbor classifier by analyzing wildfire influent attributes and wildfire consequent attributes. Moreover, real-time emergency alert generation mechanism based on temporal mining has been proposed in event of adverse wildfire conditions. Estimation of future wildfire proneness levels of a forest terrain using Holt–Winter’s forecasting model also forms an integral part of the proposed framework. Implementation results reveal that high values of accuracy, specificity, sensitivity and precision averaging to 93.97%, 92.35%, 93.01% and 91.24% are attained for determination of wildfire proneness of a forest terrain. Low values of mean absolute error (MAE), mean square error (MSE), mean absolute percentage error and root mean square error (RMSE) averaging to 0.665, 2, 11.705 and 1.405, respectively, for real-time alert generation are registered, thereby increasing the utility of the proposed framework. Wildfire proneness forecasting also yields highly accurate results with low values of MAE, MSE and RMSE averaging to 0.166667, 0.25 and 0.492799, respectively.

Feasibility of generating synthetic CT from T1-weighted MRI using a linear mixed-effects regression model

Generation of synthetic computed tomography (sCT) for magnetic resonance imaging (MRI)-only radiotherapy is emerging as a promising direction because it can eliminate the registration error and simplify clinical workflow. The goal of this study was to generate accurate sCT from standard T1-weighted MRI for brain patients. CT and MRI data of twelve patients with brain tumors were retrospectively collected. Linear mixed-effects regression models were fitted between CT and T1-weighted MRI intensities for different segments in the brain. The whole brain sCTs were generated by combining predicted segments together. Mean absolute error (MAE) between real CTs and sCTs across all patients was 71.1 ± 5.5 Hounsfield Unit (HU). Average differences in the HU values were 1.7 ± 7.1 HU (gray matter), 0.9 ± 5.1 HU (white matter), −24.7 ± 8.0 HU (cerebrospinal fluid), 76.4 ± 17.8 HU (bone), 20.9 ± 20.4 HU (fat), −69.4 ± 28.3 HU (air). A simple regression technique has been devised that is capable of producing accurate HU maps from standard T1-weighted MRI, and exceptionally low MAE values indicate accurate prediction of sCTs. Improvement is needed in segmenting MRI using a more automatic approach.

Accuracy of Intraocular Lens Power Calculation Formulae

Aim: To investigate the accuracy of seven currently-used intraocular lens (IOL) power calculation formulae. Patients and methods: A total of 429 eyes of 429 patients were retrospectively analyzed. A low-optical coherence reflectometry device (Lenstar, Haag-Streit) was used for the calculations. SRK/T, Hoffer Q, Haigis, Holladay-2, Olsen, Hill-RBF and Barrett Universal II formulae were studied. The study included patients who underwent phacoemulsification surgery with a 2.8 mm corneal cut by a single surgeon. Refraction measurement was done at the first postoperative visit, and then spherical equivalent values were recorded for each patient. Mean absolute error (MAE) was calculated as the absolute value of postoperative spherical equivalent value minus predicted error for each formula. Results: The mean age of the patients was 67.51 ± 5.75 years. Mean axial length was 23.42 ± 2.02 mm (min: 20.03, max: 32.68, median: 23.09). Mean IOL power was 21.1 ± 5 diopters (min: -1.5, max: 3.5). Postoperative spherical values were between -1.50 and +1.00 diopters. Mean absolute errors were 0.317 for SRK/T, 0.280 for Holladay-2, 0.273 for Hoffer Q, 0.264 for Haigis, 0.347 for Barrett Universal II, 0.339 for Hill-RBF, and 0.351 for Olsen. Conclusion: Hoffer Q, Holladay-2, and Haigis formulae had significantly lower MAE values than SRK/T, Barrett Universal II, Hill-RBF, and Olsen formulae. Although the Haigis formula gave the lowest MAE value, there was no significant difference with the Hoffer Q and Holladay-2 formulae.

Parameter Estimation of Signal-Dependent Random Noise in CMOS/CCD Image Sensor Based on Numerical Characteristic of Mixed Poisson Noise Samples.

Parameter estimation of Poisson-Gaussian signal-dependent random noise in the complementary metal-oxide semiconductor/charge-coupled device image sensor is a significant step in eliminating noise. The existing estimation algorithms, which are based on finding homogeneous regions, acquire the pair of the variances of noise and the intensities of every homogeneous region to fit the linear or piecewise linear curve and ascertain the noise parameters accordingly. In contrast to the existing algorithms, in this study, the Poisson noise samples of all homogeneous regions in every block image are pieced together to constitute a larger sample following the mixed Poisson noise distribution; then, the mean and variance of the mixed Poisson noise sample are deduced. Next, the mapping function among the noise parameters to be estimated—variance of Poisson-Gaussian noise and that of Gaussian noise corresponding to the stitched region in every block image—is constructed. Finally, the unbiased estimations of noise parameters are calculated from the mapping functions of all the image blocks. The experimental results confirm that the proposed method can obtain lower mean absolute error values of estimated noise parameters than the conventional ones.

A Single-Station Approach to Model Output Statistics Temperature Forecast Error Assessment

Abstract Error characteristics of model output statistics (MOS) temperature forecasts are calculated for over 200 locations around the continental United States. The forecasts are verified on a station-by-station basis for the year 2001. Error measures used include mean algebraic error (bias), mean absolute error (MAE), relative frequency of occurrence of bias and MAE values, and the daily forecast errors themselves. A case study examining the spatial and temporal evolution of MOS errors is also presented. The error characteristics presented here, together with the case study, provide a more detailed evaluation of MOS performance than may be obtained from regionally averaged error statistics. Knowledge concerning locations where MOS forecasts have large errors or biases and why those errors or biases exist is of great value to operational forecasters. Not only does such knowledge help improve their forecasts, but forecaster performance is often compared to MOS predictions. Examples of biases in MOS forecast errors are illustrated by examining two stations in detail. Significant warm and cold biases are found in maximum temperature forecasts for Los Angeles, California (LAX), and minimum temperature forecasts for Las Vegas, Nevada (LAS), respectively. MAE values for MOS temperature predictions calculated in this study suggest that coastal stations tend to have lower MAE values and lower variability in their errors, while forecasts with high MAE and error variability are more frequent in the interior of the United States. Therefore, MAE values from samples of MOS forecasts are directly proportional to the variance in the observations. Additionally, it is found that daily maximum temperature forecast errors exhibit less variability during the summer months than they do over the rest of the year, and that forecasts for any one station rarely follow a consistent temporal pattern for more than two or three consecutive days. These inconsistent error patterns indicate that forecasting temperatures based on recent trends in MOS forecast errors at an individual station is usually not a good strategy. As shown in earlier studies by other authors and demonstrated again here, MOS temperature forecasts are often inaccurate in the vicinity of strong temperature gradients, for locations affected by shallow cold air masses, or for stations in regions of anomalously warm or cold temperatures. Finally, a case study is presented examining the spatial and temporal distributions of MOS temperature forecast errors across the United States from 13 to 15 February 2001. During this period, two surges of cold arctic air moved south into the United States. In contrast to error trends at individual stations, nationwide spatial and temporal patterns of MOS forecast errors could prove to be a powerful forecasting tool. Nationwide plots of errors in MOS forecasts would be useful if made available in real time to operational forecasters.