Sequential Analysis Puns

https://journals.lww.com/americantherapeutics/Fulltext/2021/08000/Ivermectin_for_Prevention_and_Treatment_of.7.aspx

I am making this post to dump here a lot of my current observations and analysis of the video.

First off, a disclaimer:

If you hope to prove this is real to a degree that it could be used in the public/political forum as supporting evidence, you can take your mind off it. The 'Consolas Font'/'Fake aging overlay' argument is more then enough to discredit the video, leaving all the actual footage underneath easily discreditable as manufactured. Hence, the only value left in this investigation is to ascertain wether the creature(s)/technology/events depicted here are real and what their direct(specifically derived from this material) implications are if so. Arguably that is the most valuable aspect anyway.

Personally, I have little reason to believe the creature(s) and events in these videos are not real, based on both analysis of this video (part of which I've described below) as well as other external reasons. Therefore, I seek not to make a case for the veridity of the footage, but to contribute what so far I feel this group has missed.

The following takes the visual order of the shots as submitted by someone a few posts back in a compiled processed version of the video:

Aerial pursuit -- As seen from above in flight the flying-disk design is the same as the crashed one.

Crash footage -- the video is clearly filmed from below the eye-line of the upright individual. The camera shake, spacial perspective, as well as panning motion strongly indicate the video is filmed from far away with optical zoom. This combined with the filming height and the standing individual looking in the general direction of the camera but not directly into it, suggests the video was possibly filmed from inside a vehicle that arrived at the scene. Possibly the same vehicle that shot the very first ground-level footage of the flying craft or a similar team further down at a later point.

-- the crashed craft appears the size of a small bus, which is inconsistent with the general description of such craft as at least twice that size, but that is due to the perspective flattening effect of the optically zoomed camera obfuscating the perception of the distance between the injured individual on the ground and the craft.

-- Due to the static, optically-zoomed position of the camera there is virtually no paralax from sideways motion, it is therefore safe to assume the motion of the injured individual previously pointed out in other posts is an actual sign he is still alive

The increasing interest in Bayesian group sequential design is due to its potential to reinforce efficiency in clinical trials, shorten drug development time, and enhance the accuracy of statistical inference without compromising the integrity or validity of clinical trials. In a Bayesian trial, the prior information, and the trial results, as they emerge, are viewed as a continuous stream of information, in which inferences can be updated as new data become available. Bayesian sequential designs can be proposed for clinical trials with time-to-event outcomes and alpha spending functions are used to control the overall type I error rate. Alpha spending function distributes the type I error over the duration of a sequential test. Bayes factor are often adapted for decision-making at interim analyses and present Algorithms to form decision rules and to calculate power of the proposed tests. Also, a sensitivity analysis can be executed to evaluate the impact of different choices of prior parameters on choosing critical values. The power of tests, the expected event size of the proposed design, and therefore the quality of estimators can be studied through simulations and this can be compared with the frequentist group sequential design.

Sequential designs were first proposed by Armitage (1975) in which patients are recruited in pairs and the data analyzed as the results from each pair become available. Sequential designs were extended to group sequential designs, in which patients are enrolled in successive groups instead of pairs. These have been utilized for decades by the statistical and clinical trial communities. These designs allow for multiple interim analyses at the data as the clinical trial proceeds and giving the possibility of stopping the trial early due to efficacy, futility, or safety reasons. Early termination for an efficacy trial can occur when the superiority of the treatment under study is established, for futility when the establishment of a relevant treatment difference is not likely, or for safety concerns when unacceptable adverse events become evident. This design has the power to improve the efficiency of a clinical trial by reducing its duration without lowering any scientific and regulatory standards. These designs have been widely implemented in large, long-term trials, such as phase III trials of drug development for fatal diseases, particularly where the endpoints were p

I'm working on a this approach to sentiment analysis. Pass a 1 hot vector to a CNN convoluting in 1d over n-grams. Then I would like to pass that output along with the original 1 hot sentence vector to an lstm to for sentiment classification.

I'm stuck at the stop between the CNN and lstm. What I'm doing temporarily is just putting them side by side and choosing one each time.

Anyone know of a resource out there to help with this? I'm sure it's been done before.

https://www.thelancet.com/journals/landia/article/PIIS2213-8587(18)30265-1/fulltext

Findings

We identified 81 randomised controlled trials (n=53 537 participants) that reported fracture (n=42), falls (n=37), or bone mineral density (n=41). In pooled analyses, vitamin D had no effect on total fracture (36 trials; n=44 790, relative risk 1·00, 95% CI 0·93–1·07), hip fracture (20 trials; n=36 655, 1·11, 0·97–1·26), or falls (37 trials; n=34 144, 0·97, 0·93–1·02). Results were similar in randomised controlled trials of high-dose versus low-dose vitamin D and in subgroup analyses of randomised controlled trials using doses greater than 800 IU per day. In pooled analyses, there were no clinically relevant between-group differences in bone mineral density at any site (range −0·16% to 0·76% over 1–5 years). For total fracture and falls, the effect estimate lay within the futility boundary for relative risks of 15%, 10%, 7·5%, and 5% (total fracture only), suggesting that vitamin D supplementation does not reduce fractures or falls by these amounts. For hip fracture, at a 15% relative risk, the effect estimate lay between the futility boundary and the inferior boundary, meaning there is reliable evidence that vitamin D supplementation does not reduce hip fractures by this amount, but uncertainty remains as to whether it might increase hip fractures. The effect estimate lay within the futility boundary at thresholds of 0·5% for total hip, forearm, and total body bone mineral density, and 1·0% for lumbar spine and femoral neck, providing reliable evidence that vitamin D does not alter these outcomes by these amounts.

Interpretation

Our findings suggest that vitamin D supplementation does not prevent fractures or falls, or have clinically meaningful effects on bone mineral density. There were no differences between the effects of higher and lower doses of vitamin D. There is little justification to use vitamin D supplements to maintain or improve musculoskeletal health. This conclusion should be reflected in clinical guidelines.

I know that recurrent and convolutional neural networks are widely used for time series/sequential data. Are there any other commonly used algorithms designed to work with such input and achieve somewhat comparable results?

I'm writing a proposal for a mixed methods sequential exploratory research design with the taxonomy development variant model. My goal with using this method is to find out nominal variables in the qualitative (first phase) such as gender, race, income-level to use as taxonomies in the second quantitative phase. In the second quantitative phase, there will be the use of nominal variables found in the first phase and ordinal variables explored in the second phase (feelings about things on the likert scale). So, I will have qualitative data to analyze first to create taxonomies to develop what I will be researching in the quantitative phase, then I will have to analyze the quantitative data. I've researched many methods, and I have found multiple data analysis methods that would work. I am considering using a grand theory analysis method with codifying in the first phase to develop the taxonomies, and then either using the EDA and cross-tabulation method in the second phase, or the multivariate inferential method to find correlations. Either way, my goal with this research is to explore the variables and correlations that affect the phenomena in my research. I'm unsure of what methods to use, any suggestions will be greatly appreciated.

I've been tracking every review our business has recieved since it's inception. Column A is the date, Column B is the product that was reviewed, Column C-G are the star values.

Not every date is represented here. Some days we have multiple reviews, so Lines 4-8 might be for 23 January, for example. Other dates we didn't receive any reviews.

I want to make another spreadsheet that tracks our WoW/MoM review performance, so I need to find a way to translate this data into a new sheet where every row is a sequential day of the year (20 January/21 January/23 January...) and that row tallies up the total number of reviews given on this date.

Should I be using =COUNTIF, or =IF with a COUNTA inside? I'm struggling to write the correct formula here.

Thank you.