Improving Plot-Level Model of Forest Biomass: A Combined Approach Using Machine Learning with Spatial Statistics

Estimating the aboveground biomass (AGB) at the plot level plays a major role in connecting accurate single-tree AGB measurements to relatively difficult regional AGB estimates. However, AGB estimates at the plot level suffer from many uncertainties. The goal of this study is to determine whether combining machine learning with spatial statistics reduces the uncertainty of plot-level AGB estimates. To illustrate this issue, this study evaluates and compares the performance of different models for estimating plot-level forest AGB. These models include three different machine learning models [support vector machine (SVM), random forest (RF), and a radial basis function artificial neural network (RBF-ANN)], one spatial statistic model (P-BSHADE), and three combinations thereof (SVM & P-BSHADE, RF & P-BSHADE, and RBF-ANN & P-BSHADE). The results show that the root mean square error, mean absolute error, and mean relative error of all combined models are substantially smaller than those of any individual model, with the RF & P-BSHADE combined method generating the smallest values. These results indicate that a combined approach using machine learning with spatial statistics, especially the RF & P-BSHADE model, improves the accuracy of plot-level AGB models. These research results contribute to the development of accurate large-forested-landscape AGB maps.

Evolution of Food Safety Features and Volatile Profile in White Sturgeon Caviar Treated with Different Formulations of Salt and Preservatives during a Long-Term Storage Time

Caviar is a semi-preserved fish preparation in which cold storage (around 0 °C) and packaging under anaerobic conditions are fundamental to guarantee adequate safety parameters. Consumers seem to prefer caviar prepared with food salt only, but according to the needs of the different distribution channels, some preservatives are used in order to prolong its shelf life and to allow less restrictive storage conditions. Traditionally, the most common preservative was sodium tetraborate (borax), a salt that contributes to the sensory profile of caviar. However, due to its toxicity, borax has been banned in many countries, and the current trend is to reduce or eliminate its use. In this study, we evaluated the evolution of food safety parameters (pH, water activity, microbiological parameters) and the volatile profile during 14 months of storage in caviar samples treated with three different preservatives: I. exclusively NaCl, II. a mixture of borax and NaCl, and III. a mixture of organic acids and salts. Microbial presence was studied by means of plate counts; volatile organic compounds were identified on the sample headspace by means of solid phase microextraction with gas-chromatography and mass spectrometry. Results showed relevant differences among the three treatments investigated, with salt samples characterized by the highest viable counts and the greatest presence of volatile products driven by oxidative and spoilage processes, mainly occurring toward lipid and amino acids. On the contrary, the mixture of organic acids and salts showed the best response during the entire storage period. Finally, the employment of a multiparametric statistic model allowed the identification of different clusters based on the time of ripening and the preservative treatments used.

Application of Statistic Model and Backpropagation Neural Network to Analyzing and Forecasting Hydropower Dam Displacement

Horizontal displacement of Hoa Binh dam in operation phase is analyzed and then forecasted by using three methods: the multi-regression model (MTR), the Seasonal Integrated Auto-regressive Moving Average (SARIMA) model and the Back-propagation Neural Network (BPNN) model. The monitoring data of the Hoa Binh Dam in 137 monitoring periods, including horizontal displacement, time, reservoir water level and air temperature, are used for the experiments. The results indicate that all of these three methods can describe the real trend of dam deformation and achieve the required accuracy in short-term forecast up to 9 months. In addition, forecast results of BPNN have the highest stability and accuracy.

The predictive value of TIMP-2 and IGFBP7 for kidney failure and 30-day mortality after elective cardiac surgery

Acute kidney injury (AKI) is an important risk factor for chronic kidney disease, renal replacement therapy (RRT), and mortality. However, predicting AKI with currently available markers remains problematic. We assessed the predictive value of urinary tissue inhibitor of metalloprotease-2 (TIMP-2) and insulin-like growth factor-binding protein 7 (IGFBP7) regarding the need for RRT, and 30-day mortality, in elective cardiac surgery patients. In 344 elective cardiac surgery patients, we measured urinary TIMP-2 and IGFBP7 and serum creatinine at baseline and directly after surgery. Discrimination of both urinary biomarkers was assessed by the C-statistic. Model improvement for each biomarker when added to a basic model containing serum creatinine and duration of surgery was tested by the net-reclassification index (cf-NRI) and integrated discrimination index (IDI). At baseline, mean age was 66 years and 67% were men. Of all patients, 22 required RRT following surgery. IGFBP7 pre- and post-surgery and change in TIMP-2 during surgery predicted RRT with a C-statistic of about 0.80. However, a simple model including baseline serum creatinine and duration of surgery had a C-statistic of 0.92, which was improved to 0.93 upon addition of post-surgery TIMP-2 or IGFBP7, with statistically significant cf-NRIs but non-significant IDIs. Post-surgery TIMP-2 and IGFBP predicted 30-day mortality, with C-statistics of 0.74 and 0.80. In conclusion, in elective cardiac surgery patients, pre- and peri-operative clinical variables were highly discriminating about which patients required RRT after surgery. Nonetheless, in elective cardiac surgery patients, urinary TIMP-2 and IGFBP7 improved prediction of RRT and 30-day mortality post-surgery.

Modeling and simulation methods of sea clutter based on measured data

The modeling and simulation of sea clutter are important in detecting radar targets in sea backgrounds. Because the nonstationary property of sea clutter is ignored in traditional statistical models, a new method based on measured sea clutter is proposed in this paper. First, we convert the measured sea clutter data under different sea conditions [[Formula: see text]] into real amplitude [Formula: see text]. Instantaneous phase [Formula: see text] is then extracted from the coherent radar’s baseband data. Second, we select a candidate statistic model and estimate its parameters based on [Formula: see text] by utilizing maximum likelihood estimation. Finally, we generate random series [Formula: see text] using corresponding random data generator and then add instantaneous phase [Formula: see text] into [Formula: see text], i.e., [Formula: see text], to obtain simulated sea clutter series. Through a comparison of simulated sea clutter and measured sea clutter data via histogram, the validity of the proposed method is proved.

Quantifying the Uncertainty of Parameters Measured in Spontaneous Speech of Speakers With Dementia

Purpose Corpus analyses of spontaneous language fragments of varying length provide useful insights in the language change caused by brain damage, such as caused by some forms of dementia. Sample size is an important experimental parameter to consider when designing spontaneous language analyses studies. Sample length influences the confidence levels of analyses. Machine learning approaches often favor to use as much language as available, whereas language evaluation in a clinical setting is often based on truncated samples to minimize annotation labor and to limit any discomfort for participants. This article investigates, using Bayesian estimation of machine learned models, what the ideal text length should be to minimize model uncertainty. Method We use the Stanford parser to extract linguistic variables and train a statistic model to distinguish samples by speakers with no brain damage from samples by speakers with probable Alzheimer's disease. We compare the results to previously published models that used CLAN for linguistic analysis. Results The uncertainty around six individual variables and its relation to sample length are reported. The same model with linguistic variables that is used in all three experiments can predict group membership better than a model without them. One variable (concept density) is more informative when measured using the Stanford tools than when measured using CLAN. Conclusion For our corpus of German speech, the optimal sample length is found to be around 700 words long. Longer samples do not provide more information.

Analytical Descriptions of High-Tc Cuprates by Introducing Rotating Holes and a New Model to Handle Many-Body Interactions

This paper describes all the properties of high-Tc cuprates by introducing rotating holes which are created by angular momentum conservations on a two dimensional CuO2 surface, and which have a different mass from that of a normal hole due to the magnetic field energy induced by the rotation. This new particle called a macroscopic boson describes doping dependences of pseudo gap temperature and the transition temperature at which an anomaly metal phase appears. In addition, it also describes all the properties of the anomaly metal phase, using findings from our previous article [1] . Furthermore, the present paper introduces a new model to handle many-body interactions, which results in a new statistic equation. A partition function of macroscopic bosons describes all the properties of the anomaly metal phase, which sufficiently agrees with experiments. Moreover, the above-mentioned statistic equation describing many-body interactions accurately explains why high-Tc cuprates have significantly high critical temperatures, which indicates that the source of the characteristic stems from pseudo gap energy. By introducing a macroscopic boson and the new statistic model for many-body interactions, the present paper uncovered the mystery of high-Tc cuprates, which have been a challenge for many researchers. Moreover, in the present paper, pure analytical calculations are conducted. These calculations agree with experimental data which do not employ numerical calculations or fitting methods but employ many actual physical constants.

0903 Screening Of Pediatric Obstructive Sleep Apnea Using Video Monitoring

Introduction In Japan, the many of the patients are not able to access the specialized sleep medical facilities for overnight polysomnography(PSG) due to less availability and cost issues. Purpose of the study is to examine whether combination of video monitoring and other clinical examinations can reliably predict the severity of pediatric OSA compared with PSG. Methods Between April 1, 2012 and March 31, 2019, total of 175 children (3-12 years of age, boy 122, girl 53) with SDB were enrolled in this individual prospective-cohort study. In-laboratory based PSG were performed for all patients and sleep stages and respiratory events were manually scored. Video monitoring was performed during PSG. Modified video-recording test scoring system (based on Sivan et al 1996), were scored by laboratory technicians. Other clinical examinations were extracted from each PSG with ENT examinations, cephalogram, and rhinomanometry for all patient Results Multiple linear regression analyses was performed with a forward stepwise approach in which independent predictors that were significantly related to severity of OSA (AHI: 5/hr and 10/hr). Applying the multiple logistic regression analysis, the independent predictors for AHI 5/hr were ODI 3% >3/hr, rhinomanometry (NR>0.5 Pa/cm3/sec), enlargement of tonsils (Brodsky classification more than 2), two video monitoring items and total score, with an accuracy of predictive statistic model was 88.0% (sensitivity 78.3%, and specificity 93.0%). For the severity above AHI 10/hr, the independent predictors were Cephalogram parameter (Fx>84°), Oximetry (ODI 3% >5/hr) and BMI<15 with the video monitoring parameters of whole night inspiratory noise (loud) and chest retraction contribute to predict with the sensitivity 91.5%, the specificity 82.6% and the accuracy 88.0%. Conclusion Video monitor scoring parameters contributed to predict both AHI 5/hr and 10/hr with good overall sensitivity, specificity and overall accuracy compare with the combination of objective results alone. Instead of PSG, the combination of video scoring system and multiple clinical examinations could potentially provide reliable diagnostic approach for pediatric OSA with high accuracy. These results will support to establish more efficient diagnostic strategy for both patients and physicians Support N/A