
系統識別號 
U00260408201316550400 
論文名稱(中文) 
應用主成份分析超音波參數及類神經網路模式於預估胎兒體重

論文名稱(英文) 
Application of Principal Component Analysis and Artificial Neural Network on Ultrasonographic Parameters for Fetal Weight Estimation 
校院名稱 
成功大學 
系所名稱(中) 
生物醫學工程學系 
系所名稱(英) 
Department of BioMedical Engineering 
學年度 
101 
學期 
2 
出版年 
102 
研究生(中文) 
鄭月琴 
研究生(英文) 
YuehChin Cheng 
學號 
p88951018 
學位類別 
博士 
語文別 
英文 
論文頁數 
103頁 
口試委員 
指導教授鄭國順 口試委員陳中明 口試委員施東河 口試委員李宗南 口試委員孫永年 口試委員詹寶珠 口試委員張烱心 口試委員張峰銘

中文關鍵字 
主成份分析
類神經網路
兩階段法
超音波參數
預估胎兒體重

英文關鍵字 
Principle components analysis
Artificial Neural Network
twostep cluster analysis
Ultrasonographic Parameters
Fetal Weight Estimation

學科別分類 

中文摘要 
準確的估計子宮內之胎兒大小是母胎醫學及周產期照顧領域中相當重要的發展。被預估的胎兒體重，在懷孕期，是為了評估正常及異常的胎兒成長；在產前，更是提供臨床醫師評估產婦生產的方式及時機，作為分娩前處置的重要參考依據。所以準確的估計胎兒體重，不僅降低母親及胎兒的生產風險，而且減少非必要的剖腹生產所造成醫療費用的浪費，對於提升生產品質有著實質且重要的臨床價值。在台灣，臨床上使用醫用超音波量測胎兒的生物度量參數，例如：雙頂骨間徑(BPD)、枕額徑(OFD)、頭圍(HC)、腹圍(AC)、股骨長度(FL)等，經由多項式迴歸模式來推導出胎兒體重預測值，由於待解之問題日趨複雜，對於低體重、極低體重、高體重、頭圍大腹圍小、頭圍小腹圍大、胎位不正、羊水過多及羊水過少之胎兒量測，臨床上所要求之準確性提昇，此模式已逐漸不夠使用。近年來，以高解析度的醫用超音波量測胎兒的生物度量參數，經由科學化的數值分析，再應用主成份分析及類神經網路的非線性模擬分析法可得到最理想的新生兒體重預測值。
本研究的目的為設計及發展以主成份分析為分組基礎之類神經網路預測模式，求得各組最佳預估胎兒體重輸出值。計劃之特定目標包括：(1)收集並建立胎兒超音波參數資料庫及資料統計檢定；(2)以主成份分析量化胎兒體型特徵值作為分組依據；(3)胎兒體型分組之分析；(4)建立及訓練分組基礎之倒傳遞類神經網路模式來預估胎兒體重；(5)建立超音波參數誤差補償調整模式；(6)以統計分析作為方法學之比較及進行系統可行性之評估。本研究的假說為：應用多變量統計基礎理論及現代類神經網路科技，可提供客觀量化的方法來改善應用胎兒超音波參數預估新生兒體重的準確性。
本研究方法，首先收集2,127位單胞胎新生兒其產前三天內胎兒的超音波參數值(雙頂骨間徑、枕額徑、腹圍、頭圍、股骨長度等) 及出生之新生兒真實體重，以上參數值經由(1)常態分析(KolmogorovSmirnov test)檢定資料分佈形態及項目分析(item analysis)篩檢出具有鑑別力的參數；(2)相關性分析(correlation analysis)及逐步迴歸法(stepwise regression)選取與胎兒體重最具貢獻度的超音波參數作為分組依據再利用K平均演算法 (Kmeans analysis) 進行胎兒分組之分析；(3)主成份分析(Principle components analysis)量化胎兒體型特徵值作為分組依據在利用兩階段法(twostep cluster analysis) 進行胎兒體型分組之分析；(4)建立及訓練分組倒傳遞類神經網路模式(back propagation network algorithm)，並求得各組最佳預估胎兒體重輸出值；(5)以赤池信息量準則(Akaike information criterion)及最小均方誤差法(Minimum mean squared error)建立系統及參數的誤差補償模式。以胎兒體型作為網路分組之架構，將個案資料隨機分配為二組，其中1,489位作為建立及訓練倒傳遞類神經網路(比例共軛梯度演算法)學習模組，另外638位則作為驗證模組。預測準確度的評估是採用Friedman test 統計分析法。
本研究結果以胎兒體型分組為基礎之倒傳遞網路，驗證胎兒體重預估的準確度為 (n = 638，MAPE = 4.9 ± 3.5％，MAE = 149.4 ± 110.2g) 與台灣 (1) Hsieh formula 1B迴歸分析(n = 638，MAPE = 6.0 ± 4.6％，MAE = 173.2 ± 120.3g, p＜0.01)；(2) Hsieh formula 2B迴歸分析(n = 638，MAPE = 6.5 ± 7.2％，MAE = 175.1 ± 120.4g, p＜0.01)；(3)美國Hadlock 迴歸分析(n = 638，MAPE = 7.4 ± 5.3％，MAE = 224.6 ± 169.0g, p＜0.01) 及 (4)以胎兒腹圍分組為基礎之倒傳遞網路(n = 638，MAPE = 5.3 ± 4.1％，MAE = 157.9 ± 119.9g, p＜0.01)比較後，均有顯著的改善。
經由上述工程科技及原理的整合與應用發展一套準確預估胎兒體重的專家系統。本研究的重要性在於應用統計方法將數值範圍大、且變異性高的參數進行合理的分組，以改善各組參數的異質性。故以分組訓練類神經網路來預估胎兒體重的準確度，皆高於上述文獻方法之結果。研究貢獻在懷孕期間是監測胎兒生長發育的最重要指標，在產前是提供生產方式及處置的最安全選擇，可有效的降低嬰兒與產婦併發症及死亡率。

英文摘要 
For obstetrics, prenatal assessment of fetal weight and growth is paramount. Accurate estimated fetal weight (EFW) is the principal way to measure and monitor fetal growth in utero, and thereby assist the management of delivery. EFW is an important and reliable method for predicting fetal growth disorders which can result in serious antenatal, perinatal, and postnatal complications. Clinicians rely on EFW for determining appropriate delivery mode, timing, and suitable delivery procedure to reduce maternalfetal risk and medical cost. Accurate estimation of fetal weight lowers the risk of the normal spontaneous delivery (NSD). On the other hand, clinicians may encourage parents to proceed with NSD when a safe outcome is foreseeable and reduce medical costs. Currently, ultrasound (US) is the major tool for EFW. EFW methods, such as multiple regression model, are commonly used in clinical Obstetrics. Combinations of US fetal growth parameters, such as biparietal diameter (BPD), abdominal circumference (AC) and femur length (FL), along with Hsieh’s reported equations, are the major methods for the estimation of fetal weight in Taiwan. Although Hsieh’s regression methods are commonly accepted, estimation errors increase when birth weight is greater than 4000g or less than 2000g. Estimation errors are a common problem in daily practice for obstetricians so there is room for improvement of fetal weight estimation accuracy. Extreme body variations among fetal US parameters, such as macrosomia or low birth weight result in increasing error of EFW via regressive analysis. With the assistance of realtime highresolution US and principal component analysis (PCA) along with artificial neural network (ANN), a better EFW model can be achieved.
The purpose of this research project was to design and develop an automated featurebased classification method. We attempted (1) to identify fetal characteristics through ultrasound parameters (USP) sets for classifying fetuses (2) in order to develop an artificial neural network (ANN). Our hypothesis is that using PCA and ANN can improve the accuracy of EFW. This research accomplishes the following procedures: (1) collect USP data for analysis; (2) apply PCA to extract the characteristic parameters of fetal groupings; (3) apply twostep cluster analysis for the classification of fetal groupings; (4) establish a backpropagation ANN model of the classified groups for fetal weight estimation; (5) Apply Akaike information criteria (AIC) and Minimummeansquareerror (MMSE) compensation methods to USP data; (6) compare the clinical efficacy of our proposed ANN models and other conventional regression models of EFW.
The research was divided into five stages: (1) Collection of US measurements of fetal biometric parameters and preliminary parameters analysis; (2) Calculation of correlation analysis, stepwise regression and Kmeans with one parameter for fetal size classification; (3) Apply PCA and twostep cluster algorithm with multiple parameters for fetal size classification; (4) Use AIC and MMSE model to compensate for US errors; (5) Use ANN modeling on the classified groups for EFW. The dataset used in this study came from 2,127 singleton fetuses examined within 3 days prior to delivery. The cases were randomly divided into a training set with 1,489 samples and a testing set with 638 samples. The Friedman test was used for comparing the performance of the proposed ANN model with the regressionbased EFW formulas.
Our experimental results showed the accuracy of EFW from our PCAbased ANN model was significantly better than the results generated by the Taiwanese conventional regression analysis equations (i.e. Hsieh’s formula 1B, or Hsieh’s formula 2B), and even better the results using the Western conventional regression equations (all p < 0.01). The mean absolute percent error (MAPE) and the mean absolute error (MAE) was 4.9 ± 3.5% and 149.4 ± 110.2 g for our proposed PCAbased ANN model, 6.0 ± 4.6% and 173.2 ± 120.3 g for the Hsieh 1B model, 6.5 ± 7.2% and 175.1 ± 120.4 g for the Hsieh 2B model, 7.4 ± 5.3% and 224.6 ± 169.0 g for the Hadlock model , and 5.3 ± 4.1% and 157.9 ± 119.9 g for our ACbased ANN model, respectively (all p <0.01).
Our study proves that statistical selection of US parameters for grouping, together with different ANN models, can improve EFW accuracy, even in fetuses with weight at range extremities. This study also shows the following are crucial to improve EFW accuracy: (1) considering and controlling the heterogeneity using statistics among the high variability and broad ranged US parameters, and (2) using extracted parameters to classify fetuses into suitable group for each ANN model. Establishment of standardized criteria for group classification is indispensable for the subsequent cluster methods. We believe our more accurate EFW models can decrease the risks of NSD and lower medical costs by contributing to better clinical decisionmaking, and reduce maternalfetal morbidity and mortality.

論文目次 
中文摘要 I
Abstract IV
誌謝 VIII
TABLE OF CONTENTS IX
LIST OF TABLES XII
LIST OF FIGURES XIV
Chapter 1 Introduction 1
1.1 Estimation of Fetal Weight 3
1.1.1 Fetal Ultrasound Biometry for Evaluating Fetal Weight 3
1.1.2 Regression Model 8
1.1.3 Artificial Neural Network Model 11
1.2 Principal Component Analysis 13
1.3 Kmeans Method 14
1.4 Backpropagation Networks 17
1.5 Motivation and Objectives 19
1.5.1 Purpose and Specific Aims 19
1.5.2 Research Hypothesis 21
1.5.3 Significance 21
Chapter 2 Materials and Methods 22
2.1 Data Collection 24
2.1.1 Ultrasonographic Measurement of Fetal Biometric Parameters 25
2.1.2 Descriptive Statistics and Item Analysis 29
2.2 Ultrasonographic Parameters (USPs) Analysis 30
2.2.1 ACbased Group Analysis 30
2.2.2 PCAbased Group Analysis 32
2.3 ANN Development for Fetal Weight Estimation 35
2.3.1 ACbased ANN Model 35
2.3.2 PCAbased ANN Model 39
2.4 Minimum Mean Squared Error Framework for Compensating USPs 43
2.4.1 AIC Database Partition and Subset Number Determination 45
2.4.2 Minimum Mean Squared Error (MMSE) Compensation 47
2.5 Accuracy Comparison and Performance Evaluation 48
Chapter 3 Results 51
3.1 Data Description 51
3.1.1 Descriptive Statistics 51
3.1.2 Strong Correlation between USP and BW 53
3.2 Results of ACbased Grouping 56
3.3 Results of Principal Component Analysis 57
3.4 Performance of ACbased ANN Model 60
3.5 Performance of PCAbased ANN Model 66
3.6 Performance of MMSE Model Combined with AIC 74
Chapter 4 Discussion 79
4.1 Characteristics of the ACbased Grouping 80
4.2 Effects on the PCAbased Transformation 81
4.3 Comparison and Efficacy Evaluation of Modelbased Approaches 83
4.3.1 Effects of ACbased ANN Model for EFW 83
4.3.2 Effects on PCAbased ANN Model in EFW 84
4.3.3 Effects on MMSE Model Combined with AIC in EFW 86
Chapter 5 Conclusions and Recommendations 88
References 90
著作權聲明 103

參考文獻 
[1]D.D. McIntire, S.L. Bloom, B.M. Casey, and K.J. Leveno, “Birth weight in relation to morbidity and mortality among newborn infants,” N Engl J Med, vol.340, no. 16, pp.12348, Apr, 1999.
[2]E. Merz, “Ultrasound in Obstetrics and Gynecology,” New York, Thieme, chap. 14, pp. 164, 2005.
[3]F.J. Hsieh, F.M. Chang, and H.C. Huang, “Computerassisted analysis for prediction of fetal weight by ultrasoundcomparison of biparietal diameter, abdominal circumference and femur length,” J Formosan Med Assoc, vol.86, pp.95764, 1987.
[4]F.M. Chang, H.C. Ko, Y.S. Lin, B.L. Yao, C.H. Wu, P.L. Kuo and C.H. Liu, “Clinical validation of two equation in antenatal prediction of Chinese fetal weight by ultrasonography,” J Formosan Med Assoc, vol.90,pp.108692, 1991.
[5]I.M. Usta, S. Hayek, F. Yahya, A.AbuMusa, and A.H. Nassar, “Shoulder dystocia: what is the risk of recurrence?,” Acta Obstet Gynecol Scand, vol. 87, no. 10, pp.9927, 2008.
[6]F.G. Cunningham, K.J. Leveno, S.L. Bloom, J.C. Hauth, L. Gilstrap III, and K.D. Wenstrom, “Williams Obstetrics twentysecond edition,” New York, McGrawHill, chap.20,pp.514, 2005.
[7]K.H. Hsu, P.J. Liao, and C.J.Hwang, “Factors affecting Taiwanese women's choice of cesarean section,” Soc Sci Med, vol.66, no.1, pp.2019, Jan, 2008.
[8]World Health Organization, “Appropriate technology for birth,” Lancet vol.2, no.8452, pp.4367,1985.
[9]S. Campbell, and D. Wilkin, “Ultrasonic measurement of the fetal abdominal circumference in the estimation of fetal weight,” Br J Obstet Gynaecol, vol. 82, no. 9, pp. 689–697, Sep, 1975.
[10]S. L. Warsof, P. Gohari, R. L. Berkowitz, and J. C. Hobbins , “The estimation of fetal weight by computerassisted analysis,” Am J Obstet Gynecol, vol. 128, no. 8, pp. 881–892, Aug. 1977.
[11]M. J. Shepard, V. A. Richards, R. L. Berkowitz, S. L. Warsof, and J. C. Hobbins, “An evaluation of two equations for predicting fetal weight by ultrasound,” Am J Obstet Gynecol, vol. 142, no. 1, pp. 47–54, Jan. 1982.
[12]C. P. Weiner, R. E. Sabbagha, N. Vaisrub, and M. L. Socol, “Ultrasonic fetal weight prediction: role of head circumference and femur length,” Obstet Gynecol, vol. 65, no. 6, pp. 812–817, Jun. 1985.
[13]F. P. Hadlock, R. B. Harrist, R. S. Sharman, R. L. Deter, and S. K. Park, “Estimation of fetal weight with the use of head, body and femur measurementsa prospective study,” Am J Obstet Gynecol, vol. 151, no. 3, pp. 3337, Feb. 1985.
[14]B.R. Benacerraf, R. Gelman, and F.D. Jr. Frigoletto, “Sonographically estimated fetal weights: accuracy and limitation,” Am J Obstet Gynecol, vol.159, no.5, pp.111821, Nov, 1988.
[15]N. J. Dudley, “A systematic review of the ultrasound estimation of fetal weight,” Ultrasound Obstet Gynecol, vol. 25, no. 1, pp. 80–89, Jan. 2005.
[16]N. C. Hart, A. Hilbert, B. Meurer, M. Schrauder, M. Schmid, J. Siemer, M. Voigt, and R. L. Schild, “Macrosomia: a new formula for optimized fetal weight estimation,” Ultrasound Obstet Gynecol, vol. 35, no. 1, pp. 427, Jan. 2010.
[17]J. Gardosi, M. Mongelli, M. Wilcox, and A. Chang, “An adjustable fetal weight standard,” Ultrasound Obstet Gynecol, vol. 6, no.3, pp. 16874, Sep. 1995.
[18]R. E. Sabbagha, J. Minogue, R. K. Tamura, and S. A. Hungerford, “Estimation of birth weight by use of ultrasonographic formulas targeted to large, appropriate, and smallforgestationalage fetuses,” Am J Obstet Gynecol, vol. 160, no. 4, pp. 85462, Apr. 1989.
[19]J.S. Woo, C.W. Wan, and K.M. Cho, “Computerassisted evaluation of ultrasonic fetal weight prediction using multiple regression equations with and without the fetal femur length,” J Ultrasound Med, vol.4, no.2, pp.657, Feb, 1985.
[20]Bureau of Health Promotion, Department of Health, R.O.C. (Taiwan) http://www.bhp.doh.gov.tw/BHPNet/Web/HealthTopic/TopicArticle.aspx?id=201109200013&parentid=201109200006.
[21]J.A. Martin, B.E. Hamilton, S.J.Ventura, M.J.K. Osterman, E.C. Wilson, and T.J. Mathews, “Births: final data for 2010,” Natl Vital Stat Rep, vol.61, no.1, pp.172, Aug, 2012.
[22]K. Marsál, P.H. Persson, T. Larsen, H. Lilja, A. Selbing, and B. Sultan,“ Intrauterine growth curves based on ultrasonically estimated foetal weights,” Acta Paediatr, vol.85, no.7, pp.8438, Jul, 1996.
[23]D.S. Sahota, T.Y. Leung, T.N. Leung, O.K. Chan, and T.K. Lau, “Fetal crownrump length and estimation of gestational age in an ethnic Chinese population,” Ultrasound Obstet Gynecol, vol. 33, no.2, pp.157160, Feb, 2009.
[24]H.P. Robinson, “Sonar measurement of fetal crownrump length as means of assessing maturity in first trimester of pregnancy,” Br Med J, vol. 4, no.5883, pp. 2831, Oct, 1973.
[25]H.P. Robinson, and J.E. Fleming, “A critical evaluation of sonar “crownrump length” measurements,” Br J Obstet Gynecol. Vol. 82, no.9, pp.70210, Sep, 1975.
[26]A. J. Parker, P. Davies, and J. R. Newton, “Assessment of gestational age of the Asian fetus by the sonar measurement of crownrump length and biparietal diameter,” British Journal of Obstetrics & Gynaecology, vol. 89, no. 10, pp. 8368, Oct. 1982.
[27]R.E. Sabbagha, and M. Hughey, “Standardization of sonar cephalometry and gestational age,” Obstet Gynecol, vol. 52, no. 4, pp.4026, Oct, 1978.
[28]S. Campbell, and A. Thoms, “Ultrasound measurement of the fetal head to abdomen circumference ratio in the assessment of growth retardation,” Br J Obstet Gynaecol, vol. 84, no.3 pp.16574, Mar, 1977.
[29]F.P. Hadlock, R.L. Deter, R.J. Carpenter, and S.K. Park, “Estimating fetal age: effect of head shape on BPD,” AJR, vol.137, no. 1, pp.835, Jul, 1981.
[30]F. P. Hadlock, R. L. Deter , R. B. Harrist, and S. K. Park, “Fetal biparietal diameter: a critical reevaluation of the relation to menstrual age using realtime ultrasound,” J Ultrasound Med, vol. 1, no. 3, pp. 97–104, Apr. 1982.
[31]F.P. Hadlock, R.L. Deter, R.B. Harrist, and S.K. Park, “Computer Assisted analysis of fetal age in the third trimester using multiple fetal growth parameters,” J Clin Ultrasound, vol.11, no.6, pp.3136, Aug, 1983.
[32]F.P. Hadlock, R.L. Deter, R.B Harrist, and S.K. Park, “Estimating fetal age: computer assisted analysis of multiple fetal growth parameters,” Radiology, vol.152, no.2, pp.497501, 1984.
[33]R.G. Law, and K.D. MacRae, “Head circumference as an index of fetal age,” J Ultrasound Med, vol.1, no. 7, pp.2818, Sep, 1982.
[34]W.J. Ott, “The use of ultrasonic fetal head circumference for predicting expected date of confinement,” J Clin Ultrasound, vol.12, no. 7, pp.4115, Sep, 1984.
[35]C.B. Kasby, and V. Poll, “The breech head and its ultrasound significance,” Br J Obstet Gynecol, vol.89, no. 2, pp.10610, Feb, 1982.
[36]F.P. Hadlock, R.L. Deter, R.B. Harrist, and S.K. Park, “Fetal abdominal circumference as a predictor of menstrual age,” AJR Am J Roentqenol, vol.139, no. 2, pp.36770, Aug, 1982.
[37]A.B. Kurtz, R.J. Wapner, and R.J. Kurtz, “Analysis of biparietal diameter as an accurate indicator of gestational age,” J Clin Ultrasound, vol.8, no. 4, pp.31926, Aug, 1980.
[38]P. H. Persson and B. M. Weldner, “Reliability of ultrasound fetometry in estimating gestational age in the second trimester,” Acta Obstet Gynecol Scand, vol. 65, no. 5, pp. 4813, 1986.
[39]F.A. Chervenak, D.W. Skupski, R. Romero, M.K. Myers, M. SmithLevitin, Z. Rosenwaks, and H.T. Thaler, “How accurate is fetal biometry in the assessment of fetal age?,” Am J Obstet Gynecol, vol.178, no.4, pp.67887, Apr, 1998.
[40]T.N. Leung, M.W. Pang, S.S. Daljit, T.Y. Leung, C.F. Poon, S.M. Wong, and T.K. Lau, “Fetal biometry in ethnic Chinese: biparietal diameter, head circumference, abdominal circumference and femur length,” Ultrasound Obstet Gynecol, vol.31, no.3, pp.3217, Mar, 2008.
[41]D.L. Gray, G.S. Songster, C.A. Parvin, and J.P. Crane, “Cephalic index: a gestational agedependent biometric parameter,” Obstet Gynecol, vol.74, no.4, pp.6003, Oct, 1989.
[42]P. Taipale, and V. Hiilesmaa, “Predicting delivery date by ultrasound and last menstrual period in early gestation,” Obstet Gynecol, vol.97, no.2, pp.18994, Feb, 2001.
[43]D.A. Guidetti, M.Y. Divon, J.J. Braverman, O. Langer, and I.R. Merkatz, “Sonographic estimates of fetal weight in the intrauterine growth retardation population,” Am J Perinatol, vol.7, no.1, pp.57, Jan, 1990.
[44]G. Jacobsen, “Detection of intrauterine growth deviation. A comparison between symphysisfundus height and ultrasonic measurements,” Int J Technol Assess Health Care, vol.8, pp.1705, 1992.
[45]F.P. Hadlock, R.B. Harrist, R.J. Carpenter, R.L. Deter, and S.K. Park, “Sonographic estimation of fetal weight. The value of femur length in addition to head and abdomen measurements,” Radiology, vol.150, no.2, pp.53540, Feb, 1984.
[46]F.M. Chang, R.I. Liang, H.C. Ko, B.L. Yao, C.H. Chang, and C.H. Yu, “Threedimensional ultrasoundassessed fetal thigh volumetry in predicting birth weight,” Obstet Gynecol, vol.90, no.3, pp.3319, Sep, 1997.
[47]R.L. Schild, R. Fimmers, and M. Hansmann, “Fetal weight estimation by threedimensional ultrasound,” Ultrasound Obstet Gynecol, vol.16, no.5, pp.44552, Oct, 2000.
[48]R. M. Farmer, A. L. Medearis, G. I. Hirata, and L. D. Platt, “The use of a neural network for the ultrasound estimation of fetal weight in the macrosomic fetus,” Am J Obstet Gynecol, vol. 166, no.5, pp. 146772, May, 1992.
[49]L. Chuang, J. Y. Hwang, C. H. Chang, C. H. Yu, and F. M. Chang, “Ultrasound estimation of fetal weight with the use of computerized artificial neural network model,” Ultrasound Med Biol vol. 28, no. 8, pp. 9916, Aug, 2002.
[50]Richard A. Johnson and Dean W. Wichern, “Applied Multivariate Statistical Analysis sixth edition,” U.S.A., Pearson Prentice Hill, 2007,chap.8,pp.430.
[51]S.R. Veena, G.V. Krishnaveni, A.K. Wills, J.C. Hill, and C.H. Fall, “A principal components approach to parenttonewborn body composition associations in South India,” BMC Pediatr, vol.24, no.9:16, pp.111, Feb, 2009.
[52]J. B. MacQueen, “Some Methods for classification and Analysis of Multivariate Observations,” in Proceedings of 5th Berkeley Symposium on Mathematical Statistics and Probability, vol.1, pp. 28197, 1967.
[53]L. Zhang, G.W. Small, A.S. Haka, L.H. Kidder, and E.N. Lewis, “Classification of Fourier transform infrared microscopic imaging data of human breast cells by cluster analysis and artificial neural networks,” Appl Spectrosc, vol.57, no.1, pp.1422, Jan, 2003.
[54]E. Salamalekis, P. Thomopoulos, D. Giannaris, I. Salloum, G. Vasios, A. Prentza, and D. Koutsouris, “Computerised intrapartum diagnosis of fetal hypoxia based on fetal heart rate monitoring and fetal pulse oximetry recordings utilising wavelet analysis and neural networks,” BJOG, vol.109, no.10, pp.113742, Oct, 2002.
[55]L. Fausett, “Fundamentals of Neural Networks: an architectures, algorithms, and applications,” Prentice Hall ,1994.
[56]I. Etikan, and M.K.Caglar, “Prediction methods for babies' birth weight using linear and nonlinear regression analysis,” Technol Health Care, vol.13, no.2, pp.1315, 2005.
[57]S. Kol, I. Thaler, N. Paz, and O. Shmueli, “Interpretation of nonstress tests by an artificial neural network,” Am J Obstet Gynecol, vol.172, no. 5, pp.13729, May, 1995.
[58]J.M. Miller,G.A. Kissling, H.L. Brown, and H.A. Gabert, “Estimated fetal weight: applicability to small and largeforgestationalage fetus,” J Clin Ultrasound,vol.16, no.2, pp.957, Feb, 1988.
[59]Y.C. Cheng, G.L. Yan, Y.H. Chiu, F.M. Chang, C.H. Chang, K.C. Chung, “Efficient fetal size classification combined with artificial neural network for estimation of fetal weight,” Taiwan J Obstet Gynecol, vol.51, no.4, pp.54553, Dec, 2012.
[60]F.P. Hadlock, R.L. Deter, R.B. Harrist, and S.K. Park, “Fetal biparietal diameter: rational choice of plane of section for sonographic measurement,” AJR Am J Roentgenol, vol.138, no.5, pp.8714, May, 1982.
[61]S. Campbell, “An improved method of fetal cephalometry by ultrasound,” J Obstet Gynaecol Br Commonw, vol.75, no.5, pp.56876, May, 1968.
[62]C. B. Benson and P. M. Doubilet, “Sonographic prediction of gestational age: accuracy of second and thirdtrimester fetal measurements,” AJR Am J Roentgenol, vol.157, no.6, pp.12757, Dec, 1991.
[63]F.P. Hadlock, R.B. Harrist, Y.P. Shah, D.E. King, S.K. Park, and R.S. Sharman, “Estimating fetal age using multiple parameters: a prospective evaluation in a racially mixed population,” Am J Obstet Gynecol, vol.156, no.4, pp.9557, Apr, 1987.
[64]K.A. Ruvolo, R.A. Filly, and P.W. Callen, “Evaluation of fetal femur length for prediction of gestational age in a racially mixed obstetric population,” J Ultrasound Med, vol.6, no.8, pp.4179, Aug, 1987.
[65]N. Melamed, A. BenHaroush, I. Meizner, R. Mashiach, M. Glezerman, and Y. Yogev, “Accuracy of sonographic weight estimation as a function of fetal sex,” Ultrasound Obstet Gynecol, vol. 38, no. 1, pp. 6773, Jul. 2011.
[66]N. Melamed, A. BenHaroush, I. Meizner, R. Mashiach, Y. Yogev, and J. Pardo, “Accuracy of sonographic fetal weight estimation: a matter of presentation,” Ultrasound Obstet Gynecol, vol.38, no. 4, pp.41824, Oct, 2011.
[67]M. Wilcox, J. Gardosi, M. Mongelli, C. Ray, and I. Johnson, “Birth weight from pregnancies dated by ultrasonography in a multicultural British population,” BMJ, vol. 307, no. 6904, pp. 58891, Sep, 1993.
[68]R. L. Gorsuch, “Exploratory factor analysis: its role in item analysis,” Journal of Personality Assessment, vol. 68, no. 3, pp. 53260, Jun, 1997.
[69]P. W. Lei, S. B. Dunbar, and M. J. Kolen, “A Comparison of parametric and nonparametric approaches to item analysis for multiplechoice tests,” Educational and Psychological Measurement”, vol. 64, no. 4, pp. 56587, Aug, 2004.
[70]J. A. Hartigan, and M. A. Wang, “A Kmeans clustering algorithm,” Applied Statistics, vol. 28, pp. 1008, 1979.
[71] D. S. Modha, and W. S. Spangler, “Feature Weighting in kMeans Clustering,” Machine Learning, vol. 52, no. 3, pp. 21737, 2003.
[72]S. A. Billings, and W. S. F. Voon, “A predictionerror and stepwiseregression estimation algorithm for nonlinear systems,” Int J Control, vol. 44, no. 3, pp. 80322, 1986.
[73]D. A. Henderson, and D. R. Denison, “Stepwise regression in social and psychological research,” Psychological Reports, vol. 64, pp. 2517, 1989.
[74]R.R. Castro, M. Magini, S. Pedrosa, A.R. Sales, and A.C. Nóbrega, “ Principal components analysis to evaluate ventilatory variability: comparison of athletes and sedentary men,” Med Biol Eng Comput, vol.49, no.3, pp.30511, Mar, 2011.
[75]D. Rosenberg, A. Handler, and S. Furner, “A new method for classifying patterns of prenatal care utilization using cluster analysis,” Matern Child Health J, vol.8, no.1, pp.1930, Mar, 2004.
[76]P. S. La Rosa, A. Nehorai, H. Eswaran, C. L. Lowery, and H. Preissl, “Detection of uterine MMG contractions using a multiple change point estimator and the Kmeans cluster algorithm,” IEEE Trans Biomed Eng, vol. 55, no. 2, pp. 45367, Feb, 2008.
[77]W. G. Baxt, “Application of artificial neural networks to clinical medicine,” Lancet, vol. 346, no. 8983, pp. 11358, 1995.
[78]H. Akaike, “A new look at the statistical model identification,” IEEE Transactions on Automatic Control, vol. 19, pp.71623, 1974.
[79]K.P. Burnham, and D.R. Anderson, “Multimodel inference: understanding AIC and BIC in Model Selection,” Sociological Methods and Research, vol.33, no.2, pp.261304, 2004.
[80]J. Bibby, H. Toutenburg, “Prediction and Improved Estimation in Linear Models,” John Wiley & Sons Inc, 1978.
[81]S. S. Cross, R. F. Harrison, and R. L. Kennedy, “Introduction to neural networks,” Lancet, vol. 346, no.8982, pp. 10759, Oct, 1995.
[82]C. P. Chen, F. M. Chang, C. H. Chang, Y. S. Lin, C. Y. Chou , and H. C. Ko, “Prediction of fetal macrosomia by single ultrasonic fetal biometry,” J Formosan Med Assoc, vol. 92, no. 1, pp. 248, Jan, 1993.
[83]S. P. Chauhan, E. F. Magann, R. W. 3rd Naef, J. N. Jr Martin, J. C. Morrison , “Sonographic assessment of birth weight among breech presentations,” Ultrasound Obstet Gynecol, vol. 6, no.1, pp. 547, Jul,1995.
[84]R. O. Davis, G. R. Cutter, R. L. Goldenberg, H. J. Hoffman, S. P. Cliver , and C. G. Brumfield, “Fetal biparietal diameter, head circumference, abdominal circumference and femur length. A comparison by race and sex,” J Reprod Med, vol. 38, no. 3, pp. 2016, Mar, 1993.
[85]J. Siemer, A. Hilbert, T. Wolf, N. Hart, A. Müller, and R. L. Schild, “ Genderspecific weight estimation of fetuses between 2,501 and 3,999 gnew regression formulae,” Fetal Diagn Ther, vol. 24, no. 3, pp. 3049, Oct, 2008.
[86]R. L. Schild, C. Sachs, R. Fimmers, U. Gembruch and M. Hansmann, “ Sexspecific fetal weight prediction by ultrasound,” Ultrasound Obstet Gynecol, vol. 23, no. 1, pp. 305, Jan, 2004.
[87]C.P. Chen, “ Prenatal sonographic features of fetuses in trisomy 13 pregnancies (I),” Taiwan J Obstet Gynecol, vol.48, pp.2107, 2009.
[88]C.P. Chen, “Prenatal sonographic features of fetuses in trisomy 13 pregnancies (II),” Taiwan J Obstet Gynecol, vol.48, pp.21824, 2009.
[89]C.P. Chen, “Prenatal sonographic features of fetuses in trisomy 13 pregnancies (III),” Taiwan J Obstet Gynecol, vol.48, pp.3429, 2009.
[90]C.P. Chen, “Prenatal sonographic features of fetuses in trisomy 13 pregnancies (IV),” Taiwan J Obstet Gynecol, vol. 49, pp. 312, 2010.
[91]C.P. Chen, “Prenatal diagnosis and genetic counseling for mosaic trisomy 13,” Taiwan J Obstet Gynecol, vol. 49, pp.1322, 2010.
[92]C.P. Chen, “Prenatal diagnosis, fetal surgery, recurrence risk and differential diagnosis of neural tube defects,” Taiwan J Obstet Gynecol, vol.47, pp.28390, 2008.
[93]S.C. Lu, C.H. Chang, C.H. Yu, L. Kang, P.Y. Tsai, and F.M. Chang, “Reappraisal of fetal abdominal circumference in an Asian population: analysis of 50,131 records,” Taiwan J Obstet Gynecol, vol.47, pp.4956, 2008.

論文全文使用權限 
同意授權校內瀏覽/列印電子全文服務，於20180815起公開。同意授權校外瀏覽/列印電子全文服務，於20180815起公開。 


