Multiscale Decomposition of Big Data Time Series for Analysis and Prediction of Macroeconomic Data: A Recent Approach | Chapter 04 | Theory and Applications of Mathematical Science Vol. 3

The problem of the extraction of the relevant information for pre- diction purposes in a Big Data time series context is tackled. This issue is especially crucial when the forecasting activity involves macroeconomic time series, i.e. when one is mostly interested in finding leading variables and, at the same time, avoiding overfitted model structures. Unfortunately, the use of big data can cause dangerous overparametrization phenomena in the enter- tained models. In addition, two other drawbacks should be considered: firstly, humandriven handling of big data on a case-by-case basis is an impractical (and generally not viable) option and secondly, focusing solely on the raw time series might lead to suboptimal results. The presented approach deals with these problems using a twofold strategy: i) it expands the data in time scale domain, in the attempt to increase the likelihood of giving emphasis to possibly weak, relevant, signals and ii) carries out a multi-step dimension reduction procedure. The latter task is done by means of crosscorrelation functions (whose employment will be theoretically justified) and a suitable objective function.

Author(s) Details

Livio Fenga
ISTAT, Italian National Institute of Statistics, Italy.

View Books: http://bp.bookpi.org/index.php/bpi/catalog/book/146

Advertisement

Tank Level Prediction Using Kalman and Lainiotis Filters | Chapter 01 | Advances in Mathematics and Computer Science Vol. 4

Tank level knowledge is very important in many applications, as in oil tank. The liquid in the tank can be static, filling or emptying, or sloshing, resulting to uncertain knowledge of tank level. In this work the tank level is predicted using prediction algorithms based on Kalman and Lainiotis filters. Time invariant and steady state prediction algorithms for static model and filling/emptying model are implemented. Time varying prediction algorithms for sloshing and filling/emptying and sloshing models are also implemented. The prediction algorithms’ behavior is examined concluding that the obtained predictions are very close to the real tank level. The calculation burdens of the prediction algorithms are derived, determining the faster prediction algorithm for each model.

Author(s) Details

Professor N. Assimakis
General Department, National and Kapodistrian University of Athens, Greece.

Professor G. Tziallas
General Department, University of Thessaly, Greece.

Professor I. Anagnostopoulos
School of Mechanical Engineering, National Technical University of Athens, Greece.

MSc A. Polyzos
Cross Software Solutions IKE, Greece.

View Volume: http://bp.bookpi.org/index.php/bpi/catalog/book/97

Can We Predict Preeclampsia? | Chapter 12 | Current Trends in Medicine and Medical Research Vol. 4

Hypertensive  disorders  in  pregnancy  are  a  leading  cause  of  peripartum  morbidity  and  mortality. Preeclampsia is a heterogeneous maternal syndrome. Large  studies  have  pointed  out  the  association  of  impaired  spiral  artery  remodeling  at  the fetomaternal interphase in preeclampsia, but how exactly is the fetomaternal dialogue mediated and what are the biomarkers to detect the subclinical disease in various subsets of high-risk pregnancies is  still  a  challenge.  These  biomarkers  can  finally  be  used  to  diagnose  renal  function  (Kallikrein-creatinine ratio), vascular resistance (uterine artery Doppler), coagulation disorders (platelet volume, fibronectin,  prostacyclin,  thromboxane,  oxidant  stress  (lipid  peroxidase,  8-isoprostane,  antioxidants, anticardiolipin  antibodies,  homocysteine,  serum  uric  acid),  vascular  adaptation  (Placental  growth factor, Vascular endothelial growth factor, s-flt, s-eng) and markers ofplacental function and ischemia (placental  CRH,  CRH  bp,  activin,  inhibin, hCG).Post  partum  preeclampsia  can  be  predicted  by identifying the factors preventing the excretion of sodium, puerperal diuresis  and shift of intravascular fluid into the extra vascular compartment compartment(atrial natriuretic peptide in the first week after delivery,  natriuresis  and  inhibition  of  aldosterone,  angiotensin  II,  vasopressin).  Preeclampsia  is  a heterogeneous  disease.  The  late  onset  preeclampsia  at  or  near  term  has  low fetal  and  maternal morbidity. But the early onset preeclampsia (1%) of all preeclampsia has significant risks. Prediction of  risks  and  identification  of  subclinical  disease  is  mandatory.  The  majority  of  at  risk  groups  in multigravida  are  chronic  hypertension,  pregestational  and  gestational  diabetes,  age  and  multiple fetuses. Whereas, in primi only 14% have these risks. This suggests that there are multiple underlying etiologies  of  different  clinical  presentations.  A  clinical  algorithm  based  on  clinical,  biochemical  and ultrasound markers is outlined. Post partum eclampsia can be predicted and monitored with central venous  pressure  and  pulmonary  capillary  wedge  pressure.  The  maternal  syndrome  (proteinuria, edema  and  hypertension)  also  has  differences  in  time of  onset,  severity  and  organ  system involvement as highlighted in several studies. These clinical subpopulations need to be identified and preeclampsia predicted with rigorous definition of different biomarkers of different clinical phenotypes. The  future endeavors  should  be  to  identify  subclinical  disease  in  various  clinical  phenotypes  with these potential biomarkers in prospective longidunal studies.

Author(s) Details

Dr. Jayavelan Ramkumar

Department of Cardiothoracic Surgery, Sri Ramachandra Medical College and University, Chennai-600116, India.

Dr. Nidhi Sharma

Department of Obstetrics and Gynaecology, Saveetha Medical College, Saveetha University, Chennai-602105, India.

Read full article: http://bp.bookpi.org/index.php/bpi/catalog/view/43/191/339-1

View Volume: https://doi.org/10.9734/bpi/ctmmr/v4