Search results

Bayesian additive regression trees (BART) is a fully Bayesian approach to modeling with ensembles of trees. BART can uncover complex regression functions with high-dimensional regressors in a fairly automatic way and provide Bayesian quantification of the uncertainty through the posterior. However, BART assumes independent and identical distributed (i.i.d) normal errors. This strong parametric assumption can lead to misleading inference and uncertainty quantification. In this chapter we use the classic Dirichlet process mixture (DPM) mechanism to nonparametrically model the error distribution. A key strength of BART is that default prior settings work reasonably well in a variety of problems. The challenge in extending BART is to choose the parameters of the DPM so that the strengths of the standard BART approach is not lost when the errors are close to normal, but the DPM has the ability to adapt to non-normal errors.

President, Charles S. Goldman, M.P.; Chairman, Charles Bathurst, M.P.; Vice‐Presidents: Christopher Addison, M.D., M.P., Waldorf Astor, M.P., Charles Bathurst, M.P., Hilaire Belloc, Ralph D. Blumenfeld, Lord Blyth, J.P., Colonel Charles E. Cassal, V.D., F.I.C., the Bishop of Chichester, Sir Arthur H. Church, K.C.V.O., M.A., D.Sc., F.R.S., Sir Wm. Earnshaw Cooper, C.I.E., E. Crawshay‐Williams, M.P., Sir Anderson Critchett, Bart., C.V.O., F.R.C.S.E., William Ewart, M.D., F.R.C.P., Lieut.‐Colonel Sir Joseph Fayrer, Bart., M.A., M.D., Sir Alfred D. Fripp, K.C.V.O., C.B., M.B., M.S., Sir Harold Harmsworth, Bart., Arnold F. Hills, Sir Victor Horsley, M.D., F.R.C.S., F.R.S., O. Gutekunst, Sir H. Seymour King, K.C.I.E., M.A., the Duke of Manchester, P.C., Professor Sir Wm. Osler, Bart., M.D., F.R.S., Sir Gilbert Parker, D.C.L., M.P., Sir Wm. Ramsay, K.C.B., LL.D., M.D., F.R.S., Harrington Sainsbury, M.D., F.R.C.P., W. G. Savage, M.D., B.Sc., R. H. Scanes Spicer, M.D., M.R.C.S., the Hon. Lionel Walrond, M.P., Hugh Walsham, M.D., F.R.C.P., Harvey W. Wiley, M.D., Evelyn Wrench.

The foundation collection of the printed books now forming the Library of the British Museum was that of Sir Hans Sloane. This comprised about 40,000 volumes. To it was added in 1759 the Royal collection, begun in the time of Henry VII and inherited by George II from his predecessors on the throne.

THE CATEGORICAL TABLES of the S.C. which are used entirely for the sub‐division of subjects have only been criticized for two reasons, and neither is very serious. It has been said that in many cases they make the notation too cumbrous and the numbers too long; and someone has objected to these tables for including subjects which are already in the main schedules. A lengthy symbol is almost inseparable from minute classification, because it is impossible without enormously increasing the main tables to provide for the many forms and standpoints which require expressing if an attempt is to be made to get right up to the specific subject. An example of enormous expansion will be found in the uncompleted Library of Congress Classification, in which no fewer than 7,079 numbers are used for music, a subject which in the S.C. is even more fully detailed in 332 numbers. For instance, there is no place in the Congressional scheme for the viol family of instruments, in connection with which there is a very large literature, so that, in spite of its great array of numbers, it appears that it is possible to miss important headings even in the most ambitious scheme. This inflation is caused by the constant repetition of forms, localities, and other categories, which in the S.C. are expressed once and for all in separate tables, by numbers which always mean the same thing. Thus:—

GLASGOW was later by about one hundred and thirty years than some of the Scotch towns in establishing a printing press. Three hundred years ago, though Glasgow contained a University with men of great literary activity, including amongst others Zachary Boyd, there does not appear to have been sufficient printing work to induce anyone to establish a printing press. St. Andrews and Aberdeen were both notable for the books they produced, before Glasgow even attempted any printing.

THE following abstract in tabular form has been prepared by some junior members of the Islington Public Libraries staff for the use of candidates in Section V. of the Library Association Examination. It does not pretend to do more than set out the chief provisions of the various Public Libraries Acts in a clear manner, as an aid to the memorization of the principal powers and duties conferred upon library authorities. The whole of the Acts can be purchased through any bookseller for 1s. 4½d., and every student of librarianship is advised to procure them.

The Departmental Committee appointed to inquire into the use of preservatives and colouring matters in the preservation and colouring of food, have now issued their report, and the large amount of evidence which is recorded therein will be found to be of the greatest interest to those concerned in striving to obtain a pure and unsophisticated food‐supply. It is of course much to be regretted that the Committee could not see their way to recommend the prohibition of all chemical preservatives in articles of food and drink; but, apart from this want of strength, they have made certain recommendations which, if they become law, will greatly improve the character of certain classes of food. It is satisfactory to note that formaldehyde and its preparations may be absolutely prohibited in foods and drinks; but, on the other hand, it is suggested that salicylic acid may be allowed in certain proportions in food, although in all cases its presence is to be declared. The entire prohibition of preservatives in milk would be a step in the right direction, although it is difficult to see why, in view of this recommendation, boric acid should be allowed to the extent of 0·25 per cent. in cream, more especially as by another recommendation all dietetic preparations intended for the use of invalids or infants are to be entirely free from preservative chemicals; but it will be a severe shock to tho3e traders who are in the habit of using these substances to be informed that they must declare the fact of the admixture by a label attached to the containing vessel. The use of boric acid and borax only is to be permitted in butter and margarine, in proportions not exceeding 0·5 per cent. expressed as boric acid, without notification. It is suggested that the use of salts of copper in the so‐called greening of vegetables should not be allowed, but upon this recommendation the members of the Committee were not unanimous, as in a note attached to the report one member states that he does not agree with the entire exclusion of added copper to food, for the strange reason that certain foods may naturally contain traces of copper. With equal truth it can be said that certain foods may naturally contain traces of arsenic. Is the addition of arsenic therefore to be permitted? The Committee are to be congratulated upon the result of their labours, and when these recommendations become law Great Britain may be regarded as having come a little more into line— although with some apparent reluctance—with those countries who regard the purity of their food‐supplies as a matter of national importance.