Adi Ignatius
FROM THE JULY–AUGUST 2017 ISSUE
SAVE SHARE COMMENT TEXT SIZE PRINT
Public sentiment about globalization has taken a sharp turn. The election of Donald Trump, Brexit, and the rise of ultra-right parties in Europe are all signs of growing popular displeasure with the free movement of trade, capital, people, and information. Even among business leaders, doubts about the benefits of global interconnectedness surfaced during the 2008 financial meltdown and haven’t fully receded.
In “Globalization in the Age of Trump,” Pankaj Ghemawat, a professor of global strategy at NYU’s Stern School and at IESE Business School, acknowledges these shifts. But he predicts that their impact will be limited, in large part because the world was never as “flat” as many thought.
“The contrast between the mixed-to-positive data on actual international flows and the sharply negative swing in the discourse about globalization may be rooted, ironically, in the tendency of even experienced executives to greatly overestimate the intensity of international business flows,” writes Ghemawat. Moreover, his research suggests that public policy leaders “tend to underestimate the potential gains from increased globalization and to overestimate its harmful consequences.”
The once-popular vision of a globally integrated enterprise operating in a virtually borderless world has lost its hold, weakened not just by politics but by the realities of doing business in very different markets with very different dynamics and rules. Now is the time for business and political leaders to find a balance—encouraging policies that generate global prosperity at a level that democratic societies can accept.
Figure 1. The green line represents an overfitted model and the black line represents a regularized model. While the green line best follows the training data, it is too dependent on that data and it is likely to have a higher error rate on new unseen data, compared to the black line.
Figure 2. Noisy (roughly linear) data is fitted to a linear function and a polynomial function. Although the polynomial function is a perfect fit, the linear function can be expected to generalize better: if the two functions were used to extrapolate beyond the fit data, the linear function would make better predictions.
In statistics, overfitting is “the production of an analysis that corresponds too closely or exactly to a particular set of data, and may therefore fail to fit additional data or predict future observations reliably”.[1] An overfitted model is a statistical model that contains more parameters than can be justified by the data.[2] The essence of overfitting is to have unknowingly extracted some of the residual variation (i.e. the noise) as if that variation represented underlying model structure.[3]:45
Underfitting occurs when a statistical model cannot adequately capture the underlying structure of the data. An underfitted model is a model where some parameters or terms that would appear in a correctly specified model are missing.[2] Underfitting would occur, for example, when fitting a linear model to non-linear data. Such a model will tend to have poor predictive performance.
Overfitting and underfitting can occur in machine learning, in particular. In machine learning, the phenomena are sometimes called “overtraining” and “undertraining”.
The possibility of overfitting exists because the criterion used for selecting the model is not the same as the criterion used to judge the suitability of a model. For example, a model might be selected by maximizing its performance on some set of training data, and yet its suitability might be determined by its ability to perform well on unseen data; then overfitting occurs when a model begins to “memorize” training data rather than “learning” to generalize from a trend.
As an extreme example, if the number of parameters is the same as or greater than the number of observations, then a model can perfectly predict the training data simply by memorizing the data in its entirety. (For an illustration, see Figure 2.) Such a model, though, will typically fail severely when making predictions.
The potential for overfitting depends not only on the number of parameters and data but also the conformability of the model structure with the data shape, and the magnitude of model error compared to the expected level of noise or error in the data.[citation needed] Even when the fitted model does not have an excessive number of parameters, it is to be expected that the fitted relationship will appear to perform less well on a new data set than on the data set used for fitting (a phenomenon sometimes known as shrinkage).[2] In particular, the value of the coefficient of determination will shrink relative to the original data.
To lessen the chance of, or amount of, overfitting, several techniques are available (e.g. model comparison, cross-validation, regularization, early stopping, pruning, Bayesian priors, or dropout). The basis of some techniques is either (1) to explicitly penalize overly complex models or (2) to test the model’s ability to generalize by evaluating its performance on a set of data not used for training, which is assumed to approximate the typical unseen data that a model will encounter.
articulate – able to express your thoughts, arguments, and ideas clearly and effectively; writing or speech is clear and easy to understand
chatty – a chatty writing style is friendly and informal
circuitous – taking a long time to say what you really mean when you are talking or writing about something
clean – clean language or humour does not offend people, especially because it does not involve sex
conversational – a conversational style of writing or speaking is informal, like a private conversation
crisp – crisp speech or writing is clear and effective
declamatory – expressing feelings or opinions with great force
diffuse – using too many words and not easy to understand
discursive – including information that is not relevant to the main subject
economical – an economical way of speaking or writing does not use more words than are necessary
elliptical – suggesting what you mean rather than saying or writing it clearly
eloquent – expressing what you mean using clear and effective language
emphatic – making your meaning very clear because you have very strong feelings about a situation or subject
emphatically – very firmly and clearly
epigrammatic – expressing something such as a feeling or idea in a short and clever or funny way
epistolary – relating to the writing of letters
euphemistic – euphemistic expressions are used for talking about unpleasant or embarrassing subjects without mentioning the things themselves
flowery – flowery language or writing uses many complicated words that are intended to make it more attractive
fluent – expressing yourself in a clear and confident way, without seeming to make an effort
formal – correct or conservative in style, and suitable for official or serious situations or occasions
gossipy – a gossipy letter is lively and full of news about the writer of the letter and about other people
grandiloquent – expressed in extremely formal language in order to impress people, and often sounding silly because of this
idiomatic – expressing things in a way that sounds natural
inarticulate – not able to express clearly what you want to say; not spoken or pronounced clearly
incoherent – unable to express yourself clearly
informal – used about language or behaviour that is suitable for using with friends but not in formal situations
journalistic – similar in style to journalism
learned – a learned piece of writing shows great knowledge about a subject, especially an academic subject
literary – involving books or the activity of writing, reading, or studying books; relating to the kind of words that are used only in stories or poems, and not in normal writing or speech
lyric – using words to express feelings in the way that a song would
lyrical – having the qualities of music
ornate – using unusual words and complicated sentences
orotund – containing extremely formal and complicated language intended to impress people
parenthetical – not directly connected with what you are saying or writing
pejorative – a pejorative word, phrase etc expresses criticism or a bad opinion of someone or something
picturesque – picturesque language is unusual and interesting
pithy – a pithy statement or piece of writing is short and very effective
poetic – expressing ideas in a very sensitive way and with great beauty or imagination
polemical – using or supported by strong arguments
ponderous – ponderous writing or speech is serious and boring
portentous – trying to seem very serious and important, in order to impress people
prolix – using too many words and therefore boring
punchy – a punchy piece of writing such as a speech, report, or slogan is one that has a strong effect because it uses clear simple language and not many words
rambling – a rambling speech or piece of writing is long and confusing
readable – writing that is readable is clear and able to be read
rhetorical – relating to a style of speaking or writing that is effective or intended to influence people; written or spoken in a way that is impressive but is not honest
rhetorically – in a way that expects or wants no answer; using or relating to rhetoric
rough – a rough drawing or piece of writing is not completely finished
roundly– in a strong and clear way
sententious – expressing opinions about right and wrong behaviour in a way that is intended to impress people
sesquipedalian – using a lot of long words that most people do not understand
Shakespearean – using words in the way that is typical of Shakespeare’s writing
stylistic – relating to ways of creating effects, especially in language and literature
succinct – expressed in a very short but clear way
turgid – using language in a way that is complicated and difficult to understand
unprintable – used for describing writing or words that you think are offensive
vague – someone who is vague does not clearly or fully explain something
verbose – using more words than necessary, and therefore long and boring
well-turned – a well-turned phrase is one that is expressed well
wordy – using more words than are necessary, especially long or formal words