Non-authoritative relevance coding degrades classifier accuracy

June 21st, 2013

There has been considerable attention paid to the high level of disagreement between assessors on the relevance of documents, not least on this blog. This level of disagreement has been cited to argue in favour of the use of automated text analytics (or predictive coding) in e-discovery: not only do humans make mistakes, but they may make as many as or more than automated systems. But automated systems are only as good as the data used to train them, and production managers have an important choice to make in generating this training data. Should training annotations be performed by an expert, but expensive, senior attorney? Or can it be farmed out to the less expensive, but possibly less reliable, contract attorneys typically used for manual review? This choice comes down to a trade-off between cost and reliability—though ultimately reliability itself can be (at least partly) reduced to cost, too. The cost question still needs to be addressed; but Jeremy Pickens (of Catalyst) and myself have made a start on the question of reliability in our recent SIGIR paper, Assessor Disagreement and Text Classifier Accuracy.
Read the rest of this entry »

Why 95% +/- 2% makes little sense for e-discovery certification

May 25th, 2013

It is common in e-discovery protocols to see a requirement that the production be certified with a "95% +/- X%" sample (where "X%" takes on values such as "2%" or "5%"), leading to a required sample size being specified up front. (See, for instance, the ESI protocol that was recently debated in the ongoing Da Silva Moore case.) This approach, however, makes little sense, for two reasons. First, it specifies an accuracy in our measure, when what we want to specify is some minimal level of performance. And second, decisions about sample size and allocation should be delayed until after the (candidate) production is ready, when they can be made much more efficiently and effectively.
Read the rest of this entry »

What is the maximum recall in re Biomet?

April 24th, 2013

There has been a flurry of interest the past couple of days over Judge Miller's order in re Biomet. In their e-discovery process, the defendants employed a keyword filter to reduce the size of the collection, and input only the post-filtering documents to their vendor's predictive coding system, which seems to be a frequent practice at the current stage of adoption of predictive coding technology. The plaintiffs demanded that instead the defendants apply predictive coding to the full collection. Judge Miller found in favour of the defendants.
Read the rest of this entry »

Stratified sampling in e-discovery evaluation

April 18th, 2013

Point- and lower-bound confidence estimates on the completeness (or recall) of an e-discovery production are calculated by sampling documents, from both the production and the remainder of the collection (the null set). The most straightforward way to draw this sample is as a simple random sample (SRS) across the whole collection, produced and unproduced. However, the same level of accuracy can be achieved for a fraction of the review cost by using stratified sampling instead. In this post, I introduce the use of stratified sampling in the evaluation of e-discovery productions. In a later post, I will provide worked examples, illustrating the saving in review cost that can be achieved.
Read the rest of this entry »

Does automatic text classification work for low-prevalence topics?

January 26th, 2013

Readers of Ralph Losey's blog will know that he is an advocate of what he calls "multimodal" search in e-discovery; that is, a search methodology that involves a mix of human-directed keyword search, human-machine blended concept search, and machine-directed text classification (or predictive coding, in the e-discovery jargon). Meanwhile, he deprecates the alternative model of machine-driven review, in which only text classification technology is employed, and the human's sole function is to code machine-selected documents as responsive or non-responsive. The difference between the two modes can be seen most clearly in the creation of the seed set (that is, the initial training set created to bootstrap the text classification process). In multimodal review, the seed set might be taken from (a sample of) the results of active search by the human reviewer; in machine-driven review, the seed set is formed by a random sample from the collection.
Read the rest of this entry »

Why confidence intervals in e-discovery validation?

December 9th, 2012

A question that often comes up when discussing e-discovery validation protocols is, why should they be based on confidence intervals, rather than point estimates? That is, why do we say, for instance, "the production will be accepted if we have 95% confidence that its recall is greater than 60%"? Why not just say "the production will be accepted if its estimated recall is 75%"? Indeed, there have been some recent protocols that take the latter, point-estimate approach. (The ESI protocol of Global Aerospace, Inc. v. Landow Aviation, for instance, states simply that 75% recall shall be the "acceptable recall criterion", without specifying anything about confidence levels.) Answering this question requires some reflection on what we are trying to achieve with a validation protocol in the first place.
Read the rest of this entry »

The environmental consequences of SIGIR

December 4th, 2012

As it is becoming apparent that, without drastic immediate action, we are going to significantly overshoot greenhouse gas emission targets and warm the planet by an environmentally disastrous 4 to 5 degrees centigrade by the end of the century, I thought I should fulfil my long-standing promise to myself and calculate the carbon emissions generated by the annual SIGIR conference. I'm only going to consider here the emissions caused by air travel; but this is likely to be the overwhelming majority of total emissions.
Read the rest of this entry »

Statistical power of E-discovery validation

September 5th, 2012

My last post introduced the idea of the satisfiability of a post-production quality assurance protocol. We said that such a protocol is not satisfiable for a given size of the sample from the unretrieved (or null) set if the protocol were to fail the production even if the sample found no relevant documents. The reason a protocol could fail in such a circumstance is that the upper bound of the confidence interval on the missed relevant documents could still be above our threshold.
Read the rest of this entry »

Meaningful QA sample size in e-discovery

August 13th, 2012

In my last post, I examined the live-blog e-discovery production being performed by Ralph Losey, and asked what lower limit we could place on the recall of highly relevant documents with 95% confidence based on the final, quality assurance sample. The QA sample drew 1065 documents from the null set (that is, the set of documents that were not slated for production). Although none of these documents were highly relevant, this still only allows us to say with 95% confidence that no more than 0.281% of the null documents are highly relevant. Since there are 698,423 null documents, this represents an upper bound of 1962 highly relevant documents that have been missed. As only 18 highly relevant documents were found in the production, Ralph's lower-bound recall is 1%. To get this lower-bound recall up to 50%, you'd need to sample around 100,000 documents from the null set without finding any highly relevant.
Read the rest of this entry »

Quality assurance samples and prior beliefs

August 8th, 2012

Those who are following Ralph Losey's live-blogged production of material on involuntary termination from the EDRM Enron collection will know that he has reached what was to be the quality assurance step (though he has decided to do at least one more iteration of production for the sake of scientific verification). Quality assurance here involves taking a final sample from the part of the collection that is not to be produced -- what Ralph terms the "null set" -- and checking to see if any relevant documents have been missed. The outcome of this QA sample has led to an interesting discussion between Ralph and Gord Cormack on the use and meaning of confidence intervals, and how sure we can really be that (almost) no relevant documents have been missed. I've commented on the discussion at Ralph's blog; I though it would not be amiss to expand upon those comments here.
Read the rest of this entry »