How to read a research paper | search for and read papers with me | phd student advice

hello everyone and welcome or welcome back to pH D and productivity if you're a new year my name is Kiera and I'm a first year computer science PhD student I made this YouTube channel to work on my public speaking as well as share my experience and any sort of tips that I have for PhD students all over the world so if you would like to see videos about doing a PhD and staying productive then be sure to subscribe and hit the notification bells that you know when new videos are uploaded today I'm going to be sharing with you sort of my process for finding new papers and deciding whether these are going to be worth my time to read at a given time as well as take you through the process of reading a paper as well so make sure you hit the like button if you enjoyed this video and this video was decided on a poll in a Facebook group for PhD students so I was thinking I may start my own Facebook group just so that you guys could have more of an opportunity to get involved with what kind of videos I make so let me know in the comments below if that's something you guys would be interested in so we are starting off here on Google Scholar for a bit of context I'm currently working on my literature review and my PhD is in computer science or machine learning techniques to support endurance athletes such as marathon runners so what I'm working on at the moment is the injury prevention on prediction element of this work so the papers that I'm looking for are all related to the incidence of running related injuries the causes and risk factors as well as injury prevention and any work that has been done in the field of computer science so far for this so I'm starting off here on Google Scholar and I'm starting with a paper that I've already read as it's been quite useful for injury instances and now I'm going to be doing a citation tracing so looking through all of the different articles that cite this article to find more papers that will be relevant to what I want to talk about so clicking here uncited by i can see all of the different articles that cite this article and then this is sort of the first part of reading is looking through the titles of articles and seeing if these are relevant to your work so there's some here that I can see would be relevant or would not be relevant so this article here is by a barefoot running that's not really relevant because people won't be running a marathon barefoot but it might be useful to get even something out of it at some point but I think I will leave it for the moment then the impact of body mass on my biomechanics of recreational runners at this time I'm not really interested in the actual biomechanics of runners more so how if this article was the impact of body mass index on running related injuries then yes that would be relevant and this article here definitely sounds interesting not for what I'm working on right now but I think it's worth saving anyways so let me just do that I'm also using the Mendeley extension for Chrome so that means that when I come across an article that I want to download into my Mandalay I just have to click on the web importer and you know decide where I wanted to go and I already have it saved in here so that's where it goes at the moment and then I can just save it to my Mandalay instead of having to download it and have it take up space on my computer so I can go back now and then see if there's any more ultrasound of sports injuries of the musculature system nope that's not going to be relevant so this one might be relevant because it's to do with running shoes and how they you know the opinion of she shoe sales persons how that effects how they prescribe shoes so let's just take a look at this one so the second pass then so once an article has drawn in your attention just with the title the next thing to do will be to read the abstract and decide if this is in fact something that you want to read in full or even if it's something that you want to read part of I think a lot of people sort of get into the feeling that they need to read absolutely everything they come across and that's really not true because there's going to be a ton of things that you won't need to read the full paper of because you know you might be interested in what they found but you don't personally need to know the ins and outs of the methodology because it's not something you're going to employ yourself so just having a quick read through the abstract here so they're saying many runners and practitioners feel that the the main a main cause for running related injury is the prescription of running shoes that are unsuitable so that's interesting and this actually might so they're saying that this might be caused to you to marketing so that would be interesting perhaps to look through the intersection to see where they got these sort of facts from because that could be something that I would like to use so I'd probably be interested in finding these in the paper and seeing the original papers that they came from so then this study was looking to examine the beliefs of running shoe sale persons and physiotherapy students regarding the influence of running shoes and footprint a shoe on or or I so their main findings then are that sales persons believe they know more about running shoes and footprint Asian and reading related injury than students do so that's what the hype evaluates very significant and then people don't really believe that well thirty percent of students and only fit fourteen percent of sales persons believed arts training errors caused or or I'm lots of sales persons seem to believe that more expensive shoes are better preventing or I than cheaper shoes compared to students only fifteen percent believe this and sales persons would recommend uninjured runners although thirty nine percent of students he would recommend uninjured runners change their running shoe despite being satisfied with their current shoe so that's interesting as well so these could be like something to include as a side point perhaps about hair running shoes and you know the impact of that for different runners so I'm going to download this one so this here prognosis on prevention of injuries in recreational runners is something that I know immediately will be very interesting and very useful for my kind of topic so I've already downloaded this because it is our PhD thesis all about running alleged injuries and it's very very useful for my topics so I hope that just gave you a general insight into how I kind of conduct a citation tracing to look through different papers to see which ones will be relevant and just the initial stages to decide if I'll be reading a paper so so now we are going to be looking through a main paper for me in my Mendeley so this is going to be a main paper that I'll be using in my PhD research in general so this one's useful for you guys to see because this is an article that I'll be reading in full but just to show you the steps that are taken for how you want to read something like this so the first thing that obviously got my attention was running with cases a case based reasoning approach to running your best marathon so because my topic is marathon running and the use of computer science and machine learning this is definitely suitable to me so the first thing that I'm going to do is read through the abstract as usual and just highlight anything important but generally the whole abstract is important so it's usually highlighted in full but this is just to make sure that I do want to read this paper I'll be reading the abstract first of all so I can already see that this paper is a novel application of case-based reasoning to address predicting a personal best finish time for marathon runners and recommending a pacing plan so this is definitely something I'm going to run to read further so the next thing that I do is go down to the conclusion and just figure out what they actually find so they were able to find the best with a mean approach okay so there's not a huge amount of detail in the conclusion section but it's mainly that you know this was useful and for future work it should be developed further so I know already that this is something I definitely do want to continue reading and now I'm going to go and read the introduction in detail and this will give me a good insight into the problem and the motivation for this research so this is good as well because it tells me what kind of background info I would need to be including for a paper similar to this so incorporating ideas about why this is important so why runners would want to know predicted time why that's valuable to runners and similar for pacing and the important things that affect pacing as well as the different pacing strategies that exist so that's all good background info to know and to be including in a paper similar to this so then for this type of paper the best place to go next is to just look at some of the graphics and the plots and things I'd out to see if I can understand them and see what insights that gives me so that's what I'm doing next so reading looking through the plot and as well looking at the description of the figure and then looking through the algorithms as well just to see if I can get a basic understanding of their methods so looking through both of these can already give me a good understanding of how the system works and that's something that's really helpful rather than I will go through the entire method if I get to a stage where I understand that this paper is something I'm going to definitely use but for the moment it's possible to just look through basically the graphs and I'll go through and just to see to make sure do I understand what's going on then looking at the results I can see how what they did worked and how successful it was and that's something that's going to be useful to know as well this is about the prediction error for men on women using the different types of pays prediction models that they used how they sort of aggregated the cases together so they use three different methods and so that there are differences between the errors so I can see how they mentioned the mean was the best you can kind of see that that's true because the error is lower and here the profile similarity is higher so that means I can see how that works and a similar plot here so we're showing four different finish times how does the error change so you can see that the error is dutifully higher for these slower runners and similarly for the pacing the profile similarity is not as high for slower runners that's one drawback of their approach already we can see there still needs to be work because it doesn't suit the slower runners at the moment so that's slower runners is like five six hours and we can see that the difference in their personal bests how significant that was impacts prediction error and not so much profile similarity say I already feel like I got a decent understanding from these plots of the general findings and now I'm going to read through the results and discussion section as well so they just have a discussion section okay so I'm going to read through this and then I'll be able to understand more in more detail at the main findings from the paper since there wasn't as much of this in the conclusion so this is just further summarizing what I already learned from looking at the plots and the graphs yep so the Justice section sort of summarizes results as well as any potential issues with what might have impacted the results as well and I'm just going to read the evaluation section before I read the problem definition and sort of case based reasoning approach because it's better for me to have the understanding of how this paper work first and then when I really want to actually implement these ideas I will read the other two sections so I'm not going to go through all of this but just so that you have this idea now that's sort of the method that I take is to go through first the abstract to decide if this paper is definitely worth reading then I read the conclusion to get more of an understanding of their main findings of the paper and everything like that and then I go back and read the introduction to get the background information and the context for the problem and how this work is actually motivated then it will read through the you know figures and algorithms to get an understanding of how it worked and then go through the results in discussion sections and then lastly I'll go through the methods section for when I really want to understand the methods for implementation for myself so that is it for this video I really hope that it was helpful and enjoyable aim to put a fair amount of content over the next while since I am spending more time at home so if you would like to see more videos be sure to subscribe and hit the notification bow so that you know when they are uploaded thanks for watching and thanks for subscribing and I'll see you in the next video

Channel: PhD and Productivity

Views: 25,759

Rating: 4.9361181 out of 5

Keywords: research paper, how to read journal articles efficiently, how to read research papers efficiently, how to read journal articles, how to read a research paper quickly, how to read a scientific paper, phd student, google scholar, how to read a research paper, research papers in computer science, research papers explained, read papers with me, google scholar search tips, how to read research papers, how to read scientific papers, how i read research papers, lit review

Id: Y-eiG4BvO_k

Length: 19min 6sec (1146 seconds)

Published: Wed Apr 08 2020