Mahalanobis at the bus-stop [jatp]

Two bus-stops around the Indian Statistical Institute, featuring its founder, Prasanta Chandra Mahalanobis.

 

back to Kolkata [শুভ নব বর্ষ]

With high probability (!), I am off to Kolkata tonight, almost an exact year after I visited the city and the Department of Statistics of the University of Kolkata. On this occasion, I will attend a conference celebrating the 125th birthday of Mahalanobis and be staying at the Indian Statistical Institute, in another part of the city. Although I will only stay there for three days, I am looking forward meeting old friends and new faces, enjoying Bengali food and replenishing my reserves of Darjeeling tea. Shubho Nabobarsho (Happy New Year in Bengali)! As in the previous trip, I was also invited at a second conference at the Indian Association For Productivity, Quality & Reliability, on New Paradigms in Statistics for Scientific and Industrial Research, but decided against talking there for fear of being stuck for hours in the infamous Kolkata traffic.

Jubilee at the University of Calcutta

The main reason for my trip to India was taking part in the celebrations of the 75th anniversary of the Department of Statistics at the University of Calcutta and of the 100th anniversary of the birth of P.K. Bose (whom I did not know before visiting Kolkata). The Department of Statistics was created in 1941 by Mahalanobis, the very first statistics department in Asia. (Mahalanobis was also instrumental in creating the ISI in 1932. And Sankhyā in 1933.)  Fisher visited Calcutta very often and was very supportive of Mahalanobis’ efforts: in the corridor, the above picture of Fisher is displayed, with him surrounded by faculties and graduates from the Department when he came in 1941.

Although I missed the first two days of the conference (!), I enjoyed very much the exchanges I had with graduate students there, about my talk on folded MCMC and other MCMC and Bayesian issues. (With The Bayesian Choice being an easy conversational bridge-way between us as it is their Bayesian textbook.) The setting reminded me of the ISBA conference in Varanasi four years ago, with the graduate students being strongly involved and providing heavy support in the organisation, as well as eager to discuss academic and non-academic issue. (Plus offering us one evening an amazing cultural show of songs and dances.)

off to India

I am off to India today to take part in the celebration of the Platinum Jubilee of the Department of Statistics of the University of Calcutta, which was created in 1941 by Prasanta Chandra Mahalanobis. (One of the first cohort of students to complete their studies in this department was C.R. Rao.) The conference is organised by Asis Kumar Chattopadhyay whom I first met in Bangalore a few years ago and who visited Frédéric Arenou and myself last summer. This trip is quite exciting, from visiting this department to discovering Calcutta and Western Bengal, with a short stop in Darjeeling and the Himalayas foothills on the way there… Obviously, ‘Og mileage may vary in the coming days, depending on the wireless coverage. (But expect mostly pictures, anyway!)

ABC and cosmology

Two papers appeared on arXiv in the past two days with the similar theme of applying ABC-PMC [one version of which we developed with Mark Beaumont, Jean-Marie Cornuet, and Jean-Michel Marin in 2009] to cosmological problems. (As a further coincidence, I had just started refereeing yet another paper on ABC-PMC in another astronomy problem!) The first paper cosmoabc: Likelihood-free inference via Population Monte Carlo Approximate Bayesian Computation by Ishida et al. [“et al” including Ewan Cameron] proposes a Python ABC-PMC sampler with applications to galaxy clusters catalogues. The paper is primarily a description of the cosmoabc package, including code snapshots. Earlier occurrences of ABC in cosmology are found for instance in this earlier workshop, as well as in Cameron and Pettitt earlier paper. The package offers a way to evaluate the impact of a specific distance, with a 2D-graph demonstrating that the minimum [if not the range] of the simulated distances increases with the parameters getting away from the best parameter values.

“We emphasis [sic] that the choice of the distance function is a crucial step in the design of the ABC algorithm and the reader must check its properties carefully before any ABC implementation is attempted.” E.E.O. Ishida et al.

The second [by one day] paper Approximate Bayesian computation for forward modelling in cosmology by Akeret et al. also proposes a Python ABC-PMC sampler, abcpmc. With fairly similar explanations: maybe both samplers should be compared on a reference dataset. While I first thought the description of the algorithm was rather close to our version, including the choice of the empirical covariance matrix with the factor 2, it appears it is adapted from a tutorial in the Journal of Mathematical Psychology by Turner and van Zandt. One out of many tutorials and surveys on the ABC method, of which I was unaware, but which summarises the pre-2012 developments rather nicely. Except for missing Paul Fearnhead’s and Dennis Prangle’s semi-automatic Read Paper. In the abcpmc paper, the update of the covariance matrix is the one proposed by Sarah Filippi and co-authors, which includes an extra bias term for faraway particles.

“For complex data, it can be difficult or computationally expensive to calculate the distance ρ(x; y) using all the information available in x and y.” Akeret et al.

In both papers, the role of the distance is stressed as being quite important. However, the cosmoabc paper uses an L1 distance [see (2) therein] in a toy example without normalising between mean and variance, while the abcpmc paper suggests using a Mahalanobis distance that turns the d-dimensional problem into a comparison of one-dimensional projections.