Decision trees for forecasting

Recent years have seen an increasing cross-fertilization between the fields of decision analysis and forecasting. Decision-analytic models often require forecasts as inputs, and aspects of the Bayesian decision-theoretic framework underlying decision analysis have proved useful to forecasting, particularly in contexts where subjective judgemental inputs are required. This paper describes the use of decision tree analysis for forecasting and illustrates its use for corporate divisional forecasting and planning. A specialized decision-analytic technique, acts as events, is also described and illustrated to forecast a new product's earnings. Conclusions are drawn about the applicability of decision analysis for forecasting.     

Indoxyl derivatives of drug metabolites

Summary Indoxyl derivatives were detected as minor products among the urinary metabolites of two trial drugs, a benzodiazepine (GP 55 129) and a benzophenone (CGP 11 952). Their structures were elucidated by NMR and mass spectroscopy. Presumably, metabolites containing potential aldehyde functions react spontaneously with endogenous indoxyl. Such derivatives have not hitherto been encountered in drug metabolism.     

The melatonin rhythm: both a clock and a calendar

The paper briefly reviews the data which shows that the circadian production and secretion of melatonin by the pineal gland can impart both daily, i.e., clock, and seasonal, i.e., calendar, information to the organism. The paper summarizes the 3 patterns of nocturnal melatonin production that have been described. Clearly, regardless of the pattern of nocturnal melatonin production a particular species normally displays, the duration of nightime elevated melatonin is proportional to the duration of the night length. Since daylength under natural conditions changes daily the melatonin rhythm, which adjusts to the photoperiod sends time of year information to the organism. The melatonin receptors which subserve the clock message sent by the pineal gland in the form of a melatonin cycle may reside in the biological clock itself, namely, the suprachiasmatic nuclei (SCN). The melatonin receptors that mediate seasonal changes in reproductive physiology are presumably those that are located on the pars tuberalis cells of the anterior pituitary gland. Besides these receptors which likely mediate clock and calendar information, melatonin receptors have been described in other organs. Interestingly, the distribution of melatonin receptors is highly species-specific. Whereas the clock and calendar information that the melatonin cycle imparts to the organism relies on cell membrane receptors, a fact that is of some interest considering the high lipophilicity of melatonin, recent studies indicate that other functions of melatonin may require no receptor whatsoever.     

A study of macromolecular diffusion through native porcine mucus.

A diffusion chamber technique based on time-lag analysis for the estimation of effective diffusion coefficients of radiolabelled macromolecules of varying molecular weights through native mucus gel is reported. For all solutes studied, a reduction in effective diffusion coefficients was observed with a retardation of solute flux in both aqueous and mucus layers. Over the molecular weight range of solutes investigated (126-186,000 Daltons), a consistent effect of molecular weight was evident with regard to the retarding effect of mucus. No apparent or absolute molecular weight cut-off for macromolecular transfer was exhibited. However, at high molecular weights (greater than 30,000 Daltons) the retardation was greatly enhanced. The results confirm that mucus can be regarded as a gel with finite pores, but that it does not constitute an absolute barrier to even high molecular weight solutes.     

Newton's Easy Quadratures ``Omitted for the Sake of Brevity'

In the 1687 Principia, Newton gave a solution to the direct problem (given the orbit and center of force, find the central force) for a conic-section with a focal center of force (answer: a reciprocal square force) and for a spiral orbit with a polar center of force (answer: a reciprocal cube force). He did not, however, give solutions for the two corresponding inverse problems (given the force and center of force, find the orbit). He gave a cryptic solution to the inverse problem of a reciprocal cube force, but offered no solution for the reciprocal square force. Some take this omission as an indication that Newton could not solve the reciprocal square, for, they ask, why else would he not select this important problem? Others claim that ``it is child's play' for him, as evidenced by his 1671 catalogue of quadratures (tables of integrals). The answer to that question is obscured for all who attempt to work through Newton's published solution of the reciprocal cube force because it is done in the synthetic geometric style of the 1687 Principia rather than in the analytic algebraic style that Newton employed until 1671. In response to a request from David Gregory in 1694, however, Newton produced an analytic version of the body of the proof, but one which still had a geometric conclusion. Newton's charge is to find both ``the orbit' and ``the time in orbit.' In the determination of the dependence of the time on orbital position, t(r), Newton evaluated an integral of the form ∫dx/x <sup>n</sup> to calculate a finite algebraic equation for the area swept out as a function of the radius, but he did not write out the analytic expression for time t = t(r), even though he knew that the time t is proportional to that area. In the determination of the orbit, θ (r), Newton obtained an integral of the form ∫dx/√(1−x<sup>2</sup>) for the area that is proportional to the angle θ, an integral he had shown in his 1669 On Analysis by Infinite Equations to be equal to the arcsin(x). Since the solution must therefore contain a transcendental function, he knew that a finite algebraic solution for θ=θ(r) did not exist for ``the orbit' as it had for ``the time in orbit.' In contrast to these two solutions for the inverse cube force, however, it is not possible in the inverse square solution to generate a finite algebraic expression for either ``the orbit' or ``the time in orbit.' In fact, in Lemma 28, Newton offers a demonstration that the area of an ellipse cannot be given by a finite equation. I claim that the limitation of Lemma 28 forces Newton to reject the inverse square force as an example and to choose instead the reciprocal cube force as his example in Proposition 41. (Received August 14, 2002) Published online March 26, 2003 Communicated by G. Smith     

MeV Pb离子在非晶Si中的射程离散──离子束混合和增强扩散的影响

１．０ＭｅＶ２０８Ｐｂ离子在非晶Ｓｉ中的投影射程ＲＰ和射程偏差ΔＲＰ作为注量和温度二者的函数用背散射法进行测定．注量的变化范围为５×１０１３～７×１０１４ｃｍ－２．注入是在室温和ｔ＝－１２０℃下完成的．由由实验所确定的投影射程，射程偏差与注量或温度无关，并且分别等于２９５和７２．２ｎｍ．与ＴＲＩＭ８６的计算结果相比较，发现ＲＰ的偏离为１８％，而ΔＲＰ的偏离为３６％．ＲＰ和ΔＲＰ二者与注量及温度的无关性，排除了所观察到的与ＴＲＩＭ的矛盾是由于注入期间辐射增强扩散或离子束混合效应而引起的解释。     

Aspirin-like drugs may block pain independently of prostaglandin synthesis inhibition

Summary Using flurbiprofen, a chiral anti-inflammatory and analgesic 2-arylpropionic acid derivative, the enantiomers of which are not converted to each other (less than 5%) in rats or man, we obtained evidence that prostaglandin synthesis inhibition is primarily mediating the anti-inflammatory activity but prostaglandin synthesis independent mechanisms contribute to the analgesic effects. Thus, the S-form inhibited prostaglandin synthesis, inflammation and nociception in rats. The R-form had much less effect on prostaglandin synthesis and did not affect inflammation. It did, however, block nociception in rats almost as potently as the S-form. S-flurbiprofen, in contrast to the R-form, was clearly ulcerogenic in the gastrointestinal mucosa. These results indicate additional molecular mechanisms of analgesia and suggest the use of R-arylpropionic acids as analgesics.     

Non-Minimum Phase Nonlinear System Predictive Control Based on Local Recurrent Neural Networks

Abstract: After a recursive multi-step-ahead predictor for nonlinear systems based on local recurrent neural networks isintroduced, an intelligent PID controller is adopted to correct the errors including identified model errors and accumulatederrors produced in the recursive process. Characterized by predictive control, this method can achieve a good controlaccuracy and has good robustness. A simulation study shows that this control algorithm is very effective.