IBM Research today opened its second research location on the African continent and announced several new project collaborations in the areas of data driven healthcare, digital urban ecosystems and astronomy.

The Weather Company, an IBM Business and Rogers Media announced today an agreement to provide weather information and content customized specifically for the Canadian market. With this agreement, The Weather Company will provide in-depth weather data and forecasts, as well as curated content, across Rogers Media properties. In turn, Rogers Media will provide locally relevant articles, photos and video content across the Canadian versions of The Weather Channel app and website (weather.com) to provide residents with the most pertinent information possible. The Weather Company and Rogers Media will also align to enhance advertising across both companies’ properties, with Rogers leveraging its ad sales capabilities to monetize The Weather Channel properties in Canada.

BRITONS saw temperatures climb yesterday into the 20's as many welcomed the early Bank Holiday sunshine, however this weekend a dramatic change will take place as cold Arctic air will sweep snow in.

A HOT and sunny Saturday is set across the UK before a cooler Sunday.

THE FREEZING plunge in temperatures over the next 24 hours even sparked disbelief from the meteorologist on BBC Breakfast, who warned that Britain could go from its warmest day of the year on Saturday to being hit by snow on Sunday.

Invitation Only Research Event

Event participants

Professor Tim Benton, Research Director, Emerging Risks; Director, Energy, Environment and Resources Programme, Chatham House
Professor Paul Stevens, Distinguished Fellow, Energy, Environment and Resources Programme, Chatham House
Antony Froggatt, Senior Research Fellow, Energy, Environment and Resources Programme, Chatham House
Chair: Glada Lahn, Senior Research Fellow, Energy, Environment and Resources Programme

No one can predict who will commit a crime but in some cities math is helping detect areas where crimes have the greatest chance of occurring. Police then increase patrols in these "hot spots" in order to prevent crime. This innovative practice, called predictive policing, is based on large amounts of data collected from previous crimes, but it involves more than just maps and push pins. Predictive policing identifies hot spots by using algorithms similar to those used to predict aftershocks after major earthquakes. Just as aftershocks are more likely near a recent earthquake.s epicenter, so too are crimes, as criminals do indeed return to, or very close to, the scene of a crime. Cities employing this approach have seen crime rates drop and studies are underway to measure predictive policing.s part in that drop. One fact that has been determined concerns the nature of hot spots. Researchers using partial differential equations and bifurcation theory have discovered two types of hot spots, which respond quite differently to increased patrols. One type will shift to another area of the city while the other will disappear entirely. Unfortunately the two appear the same on the surface, so mathematicians and others are working to help police find ways to differentiate between the two so as to best allocate their resources.

The projected private flat supply for the next three to four years is 95,000 units, 2,000 more than the previous estimate.

The Transport & Housing Bureau today said there were 10,000 unsold units in completed projects at the end of March.

There were 60,000 units under construction, excluding those pre-sold by developers, and 25,000 units from disposed sites where construction can start any time.

The number of flats under construction in the first quarter was 900, while the number of units completed for the period was 4,200.

Click here for details.

Hong Kong Observatory Director Cheng Cho-ming today said he expects four to seven tropical cyclones to hit Hong Kong this year.

Presenting the weather outlook at a press briefing in the afternoon, Mr Cheng also said the overall mean temperature for the year will be higher than normal.

“The way that we produce our temperature forecast is based on a couple of things. One is based on the worldwide climate model - what they forecast for this year - and also based on some objective indicators.

“Based on all these indicators we expect that the overall mean temperature this year will be above normal, and very likely, will have a high chance of reaching the top 10 positions.”

Reviewing last year’s weather, Mr Cheng noted Hong Kong experienced its second warmest winter since records began, with a mean temperature of 18.7 degrees Celsius.

There was only one day when the temperature fell below 12 degrees Celsius.

Additionally, Mr Cheng said the new Hong Kong Hiking Trail Weather Service webpage has been launched to provide weather information services to people taking part in outdoor activities.

He added the Observatory also collaborated with the Agriculture, Fisheries & Conservation Department to provide weather information at popular hiking hotspots via QR codes.

(DOE/Oak Ridge National Laboratory) Using ORNL's now-decommissioned Titan supercomputer, a team of researchers estimated the combined consequences of many different extreme climate events at the county level, a unique approach that provided unprecedented regional and national climate projections that identified the areas most likely to face climate-related challenges.

BRIDGETOWN, Barbados, CMC – A few weeks before the official start of the 2020 Atlantic Hurricane Season, forecasters at the US-based Colorado State University are warning that the six-month period will be more active than normal. The CSU...

Invitation Only Research Event

Invitation Only Research Event

Event participants

Professor Tim Benton, Research Director, Emerging Risks; Director, Energy, Environment and Resources Programme, Chatham House
Professor Paul Stevens, Distinguished Fellow, Energy, Environment and Resources Programme, Chatham House
Antony Froggatt, Senior Research Fellow, Energy, Environment and Resources Programme, Chatham House
Chair: Glada Lahn, Senior Research Fellow, Energy, Environment and Resources Programme

The very first case of corona-virus illness was recorded on 30 January 2020, in India and the number of infected cases, including the death toll, continues to rise. In this paper, we present short-term forecasts of COVID-19 for 28 Indian states and five union territories using real-time data from 30 January to 21 April 2020. Applying Holt's second-order exponential smoothing method and autoregressive integrated moving average (ARIMA) model, we generate 10-day ahead forecasts of the likely number of infected cases and deaths in India for 22 April to 1 May 2020. Our results show that the number of cumulative cases in India will rise to 36335.63 [PI 95% (30884.56, 42918.87)], concurrently the number of deaths may increase to 1099.38 [PI 95% (959.77, 1553.76)] by 1 May 2020. Further, we have divided the country into severity zones based on the cumulative cases. According to this analysis, Maharashtra is likely to be the most affected states with around 9787.24 [PI 95% (6949.81, 13757.06)] cumulative cases by 1 May 2020. However, Kerala and Karnataka are likely to shift from the red zone (i.e. highly affected) to the lesser affected region. On the other hand, Gujarat and Madhya Pradesh will move to the red zone. These results mark the states where lockdown by 3 May 2020, can be loosened.

Cooperation between different data owners may lead to an improvement in forecast quality - for instance by benefiting from spatial-temporal dependencies in geographically distributed time series. Due to business competitive factors and personal data protection questions, said data owners might be unwilling to share their data, which increases the interest in collaborative privacy-preserving forecasting. This paper analyses the state-of-the-art and unveils several shortcomings of existing methods in guaranteeing data privacy when employing Vector Autoregressive (VAR) models. The paper also provides mathematical proofs and numerical analysis to evaluate existing privacy-preserving methods, dividing them into three groups: data transformation, secure multi-party computations, and decomposition methods. The analysis shows that state-of-the-art techniques have limitations in preserving data privacy, such as a trade-off between privacy and forecasting accuracy, while the original data in iterative model fitting processes, in which intermediate results are shared, can be inferred after some iterations.

In the field of numerical weather prediction (NWP), the probabilistic distribution of the future state of the atmosphere is sampled with Monte-Carlo-like simulations, called ensembles. These ensembles have deficiencies (such as conditional biases) that can be corrected thanks to statistical post-processing methods. Several ensembles exist and may be corrected with different statistiscal methods. A further step is to combine these raw or post-processed ensembles. The theory of prediction with expert advice allows us to build combination algorithms with theoretical guarantees on the forecast performance. This article adapts this theory to the case of probabilistic forecasts issued as step-wise cumulative distribution functions (CDF). The theory is applied to wind speed forecasting, by combining several raw or post-processed ensembles, considered as CDFs. The second goal of this study is to explore the use of two forecast performance criteria: the Continous ranked probability score (CRPS) and the Jolliffe-Primo test. Comparing the results obtained with both criteria leads to reconsidering the usual way to build skillful probabilistic forecasts, based on the minimization of the CRPS. Minimizing the CRPS does not necessarily produce reliable forecasts according to the Jolliffe-Primo test. The Jolliffe-Primo test generally selects reliable forecasts, but could lead to issuing suboptimal forecasts in terms of CRPS. It is proposed to use both criterion to achieve reliable and skillful probabilistic forecasts.

Yicheng Li, Adrian E. Raftery.

Source: The Annals of Applied Statistics, Volume 14, Number 1, 381--408.

Abstract:
Smoking is one of the leading preventable threats to human health and a major risk factor for lung cancer, upper aerodigestive cancer and chronic obstructive pulmonary disease. Estimating and forecasting the smoking attributable fraction (SAF) of mortality can yield insights into smoking epidemics and also provide a basis for more accurate mortality and life expectancy projection. Peto et al. ( Lancet 339 (1992) 1268–1278) proposed a method to estimate the SAF using the lung cancer mortality rate as an indicator of exposure to smoking in the population of interest. Here, we use the same method to estimate the all-age SAF (ASAF) for both genders for over 60 countries. We document a strong and cross-nationally consistent pattern of the evolution of the SAF over time. We use this as the basis for a new Bayesian hierarchical model to project future male and female ASAF from over 60 countries simultaneously. This gives forecasts as well as predictive distributions that can be used to find uncertainty intervals for any quantity of interest. We assess the model using out-of-sample predictive validation and find that it provides good forecasts and well-calibrated forecast intervals, comparing favorably with other methods.

Seth Flaxman, Michael Chirico, Pau Pereira, Charles Loeffler.

Source: The Annals of Applied Statistics, Volume 13, Number 4, 2564--2585.

Abstract:
We propose a generic spatiotemporal event forecasting method which we developed for the National Institute of Justice’s (NIJ) Real-Time Crime Forecasting Challenge (National Institute of Justice (2017)). Our method is a spatiotemporal forecasting model combining scalable randomized Reproducing Kernel Hilbert Space (RKHS) methods for approximating Gaussian processes with autoregressive smoothing kernels in a regularized supervised learning framework. While the smoothing kernels capture the two main approaches in current use in the field of crime forecasting, kernel density estimation (KDE) and self-exciting point process (SEPP) models, the RKHS component of the model can be understood as an approximation to the popular log-Gaussian Cox Process model. For inference, we discretize the spatiotemporal point pattern and learn a log-intensity function using the Poisson likelihood and highly efficient gradient-based optimization methods. Model hyperparameters including quality of RKHS approximation, spatial and temporal kernel lengthscales, number of autoregressive lags and bandwidths for smoothing kernels as well as cell shape, size and rotation, were learned using cross validation. Resulting predictions significantly exceeded baseline KDE estimates and SEPP models for sparse events.

Lisa Schlosser, Torsten Hothorn, Reto Stauffer, Achim Zeileis.

Source: The Annals of Applied Statistics, Volume 13, Number 3, 1564--1589.

Abstract:
To obtain a probabilistic model for a dependent variable based on some set of explanatory variables, a distributional approach is often adopted where the parameters of the distribution are linked to regressors. In many classical models this only captures the location of the distribution but over the last decade there has been increasing interest in distributional regression approaches modeling all parameters including location, scale and shape. Notably, so-called nonhomogeneous Gaussian regression (NGR) models both mean and variance of a Gaussian response and is particularly popular in weather forecasting. Moreover, generalized additive models for location, scale and shape (GAMLSS) provide a framework where each distribution parameter is modeled separately capturing smooth linear or nonlinear effects. However, when variable selection is required and/or there are nonsmooth dependencies or interactions (especially unknown or of high-order), it is challenging to establish a good GAMLSS. A natural alternative in these situations would be the application of regression trees or random forests but, so far, no general distributional framework is available for these. Therefore, a framework for distributional regression trees and forests is proposed that blends regression trees and random forests with classical distributions from the GAMLSS framework as well as their censored or truncated counterparts. To illustrate these novel approaches in practice, they are employed to obtain probabilistic precipitation forecasts at numerous sites in a mountainous region (Tyrol, Austria) based on a large number of numerical weather prediction quantities. It is shown that the novel distributional regression forests automatically select variables and interactions, performing on par or often even better than GAMLSS specified either through prior meteorological knowledge or a computationally more demanding boosting approach.

Ahmed Aziz Ezzat, Mikyoung Jun, Yu Ding.

Source: The Annals of Applied Statistics, Volume 13, Number 3, 1484--1510.

Abstract:
Accurate short-term forecasts are indispensable for the integration of wind energy in power grids. On a wind farm, local wind conditions exhibit sizeable variations at a fine temporal resolution. Existing statistical models may capture the in-sample variations in wind behavior, but are often shortsighted to those occurring in the near future, that is, in the forecast horizon. The calibrated regime-switching method proposed in this paper introduces an action of regime dependent calibration on the predictand (here the wind speed variable), which helps correct the bias resulting from out-of-sample variations in wind behavior. This is achieved by modeling the calibration as a function of two elements: the wind regime at the time of the forecast (and the calibration is therefore regime dependent), and the runlength, which is the time elapsed since the last observed regime change. In addition to regime-switching dynamics, the proposed model also accounts for other features of wind fields: spatio-temporal dependencies, transport effect of wind and nonstationarity. Using one year of turbine-specific wind data, we show that the calibrated regime-switching method can offer a wide margin of improvement over existing forecasting methods in terms of both wind speed and power.

Guillaume Kon Kam King, Antonio Canale, Matteo Ruggiero.

Source: Bayesian Analysis, Volume 14, Number 4, 1121--1141.

Abstract:
Motivated by the problem of forecasting demand and offer curves, we introduce a class of nonparametric dynamic models with locally-autoregressive behaviour, and provide a full inferential strategy for forecasting time series of piecewise-constant non-decreasing functions over arbitrary time horizons. The model is induced by a non Markovian system of interacting particles whose evolution is governed by a resampling step and a drift mechanism. The former is based on a global interaction and accounts for the volatility of the functional time series, while the latter is determined by a neighbourhood-based interaction with the past curves and accounts for local trend behaviours, separating these from pure noise. We discuss the implementation of the model for functional forecasting by combining a population Monte Carlo and a semi-automatic learning approach to approximate Bayesian computation which require limited tuning. We validate the inference method with a simulation study, and carry out predictive inference on a real dataset on the Italian natural gas market.

Winds on WASP-76b blow gaseous iron into cooler regions, where it condenses and falls to the planet’s surface as liquid

'The break in the scientific record is probably unprecedented,' one ecologist says

Bank of Finland Research Discussion Papers by Gonçalo Faria and Fabio Verona

[Jessica Curtis and Adam Hillman, both Forecasting Advisors at SAS, were co-authors of this post] The world has been dramatically impacted by the recent COVID-19 pandemic. Many of us are juggling a completely new lifestyle that was forced upon us overnight. As consumers find their way to a new normal, [...]

Retail forecasting through a pandemic was published on SAS Voices by Brittany Bullard

During FY12-17, a cumulative generation capacity of 99,209 MW was added against a target of 88,537 MW, outpacing the growth in demand and resulting in a declining trend of PLFs.

The International Journal of Forecasting has published its 2020-Q1 issue, guest edited by Spyros Makridakis and Fotios Petropoulos, and dedicated entirely to results and commentary on the M4 Forecasting Competition. This issue should be of great interest and value to business forecasting practitioners, and you get online access to it [...]

The post M4 Forecasting Competition: Results and Commentary appeared first on The Business Forecasting Deal.

With 2018's M4 Forecasting Competition behind us (although analysis, interpretation, and debate continue), the new M5 Competition starts March 2. Running through June 30, M5 is utilizing actual data provided by Walmart. It will be implemented using Kaggle's Platform, with $100,000 in prize money. Forecasting practitioners are encouraged to participate, [...]

The post Get ready for M5 with launch of the Makridakis Open Forecasting Centre appeared first on The Business Forecasting Deal.

What if you suspect something is wrong with your forecasting process? What if the process is consuming too much time and too many resources, while still delivering unsatisfactory results (lousy forecasts). What can you do about it? This post looks at the first two steps to effecting meaningful forecasting process [...]

The post 2 first steps to effecting forecasting process change appeared first on The Business Forecasting Deal.

Two weeks ago we looked at the first two steps in effecting forecasting process change: Justify your suspicions with data Communicate your findings That was the easy part. So why is it that so many organization realize they have a forecasting problem, yet are unable to do anything about it? [...]

The post The difficult step in effecting forecasting process change (1 of 2) appeared first on The Business Forecasting Deal.

Fildes and Goodwin (F&G) observed the subject (the regional subsidiary of a pharmaceutical company) was using a statistical forecasting system, but not fully trusting its output. Forecasters were making overrides to the system generated forecast to make it look like what they believed it should (e.g., following a life-cycle curve [...]

The post The difficult step in effecting forecasting process change (2 of 2) appeared first on The Business Forecasting Deal.

Forecasting During Chaos The Institute of Business Forecasting has produced an 80-minute virtual town hall on "Forecasting & Planning During the Chaos of a Global Pandemic." The on-demand video recording is available now and well worth a look. There is much solid practical guidance from an experienced panel: Eric Wilson, [...]

The post Forecasting during chaos: notes from an IBF virtual town hall appeared first on The Business Forecasting Deal.

Forecasting is a daunting task during normal conditions, and even more so during a disruption. But in times of greatest stress our smartest and most creative people stand out, and our true leaders emerge. You'll find these kinds of leaders among my colleagues at SAS -- smart and creative people [...]

The post Forecasting during chaos: predicting impact on demand and supply appeared first on The Business Forecasting Deal.