2 Burst results for "Southwest Johnson University"

"southwest johnson university" Discussed on Data Skeptic

Data Skeptic

06:43 min | 1 year ago

"southwest johnson university" Discussed on Data Skeptic

"And for good reason as are its siblings. Arena and seasonal remo or serena. These are widely deployed techniques sometimes blindly deployed techniques and. There's good empirical evidence to be using these techniques however when you investigate these things a couple of other perspectives like spectral analysis linear system theory and digital filtering. They take on a whole new light. There's a conversion process that data will be put through by these techniques in order to make it a stationary time series in other words difference in out any trend or seasonality to make it stationary. However that process can actually be a little bit lossing. In this episode. I talked with chung show lee one of the authors of the paper. Why are reema and serena not sufficient. My name is chung show carney associated professor and the southwest johnson university in trump province in china. And can you tell me a little bit about your specific research areas. What do you study my research areas. Congress daytime my machine learning and data analytics gender most specifically focused on forecasting demand focusing in retail and time series focused sich of sees. So the main pay for. I asked you on to discuss. Today is wire arena and serena or s arema not sufficient. You'd mentioned you have a good background in machine learning. I don't necessarily think of a reema as a machine learning technique. How do these two areas fit together in your mind. Actually because the site focused teams all problems and can be served by machinery and when the approached this focusing problem with fines and attorney time service models are very important solutions to forecast team problems. Other side focused is very important in today's areas because you're low many many data so always Focused in problem. We find penser is very important and we also find iron man. Sarah map and armagh. Those are very classic. Run divided news time service motives and when we do couldn't this i remind saruman model. We're fans than actually the classical extre nation or classical. Modern for iran. serena is northern sufficient sarichichekli. Either way actually approach. I remember sarim from elisa angle which is spectral lenzi's digital delivery and in your system theory so we use elisa angle to do countries a romance. I remember motive defines onto loads dench a sufficient from the rich porno view. So this is the whole ground actually starts from computer science. And i do can they as focused in problem because forecasting problem is very potent. So in new concerns i remember i remember those from another this whole idea and the hope and ground of this paper. So rima and essary remodels models. Have i believe some sort of guarantees and certainly a long history in the literature. What is attractive about techniques. Like these sinks. The iron makes a very important contribution on the so-called lung stationary time series forecasting because they can solve those non stationary time series models using the difference. Operator because as you know. I remember is the rich very beautiful that you got only approach. The coyote show the stationary panzer is focusing program on the other side in the real practice. The problem faced usually are lung station so how to transform. Those non station returned to risk problem to stationers. Tides hampshire is very important probe and i remind us over seven point elegant way to do this job. So he's attracting. This either reason in my opinion but lee finds the transformation i remote thus is not sufficient and actually we can do more because we're essentially the differentiation operator is actually can be expanded as a digital poetry than allows frequency components and realize no from cuisine components for the dior difference operator. So let's see our findings so that digital filtering on one hand. I guess maybe naively. I could look at that and say oh. It's probably good. It's taking out noise in the data. How is that the rung intuition because this regarding lahser you shoot when your candidate. Times data on the traditional explanation is voiced that pinpoint you got observation from the signal precision. Porno view is actually the pan domain. But actually you can do the cancer steak from the franck was doman so you feel with the boys. They exist though from quiz components and hybrid components the job of digital richer. Richer is to remove those from zero high from crusade. Depends on your design. Yes he remove those laws so this is a lesson as you point out in the paper. These methods end up distorting the data at certain frequencies in those are desirable frequencies that you'd wanna have so that's kind of a weakness or an achilles heel in the model. What was next once you had this observation. Yes actually our paper. We just to propose a very gender macer which is called. I remember same. Although i will do spectral lance and digital delivery and then to plow the focused. Let's he's quite simple from work i do. Is i for the spectral orleans We should develop some data dependent digital retrieve such this ritually. Let's furniture universities is a difference operator. Career affects the characteristics of the data. Which means this kind really removes alois. Because the noise kyle resistant different forms. Sometimes it's not easy for you until you find it. For example in some data points ladies annoys is low frequencies sometimes lowest. Be the high from so state dependent operator the freezer. So let's see the next job. We're going to do so. I'd love to dive deeper into the spectral analysis. I think you did a good job describing how you know noise can be in different parts of the frequency. What is it about the army sinn model. that's able to capture that. Are you actually modeling out the noise in some way or what's the next step there and actually did not apply this whole moldy are specific programs. We just was unspiritual. Allowances are electoral by going to some find. Some are real world problems and apply or miscarriage too series..

serena chung reema southwest johnson university penser Sarah map sarim elisa angle lenzi carney lee saruman Arena rima elisa Congress china iran doman hampshire
Why ARiMA Is Not Sufficient

Data Skeptic

01:50 min | 1 year ago

Why ARiMA Is Not Sufficient

"Name is chung show carney associated professor and the southwest johnson university in trump province in china. And can you tell me a little bit about your specific research areas. What do you study my research areas. Congress daytime my machine learning and data analytics gender most specifically focused on forecasting demand focusing in retail and time series focused sich of sees. So the main pay for. I asked you on to discuss. Today is wire arena and serena or s arema not sufficient. You'd mentioned you have a good background in machine learning. I don't necessarily think of a reema as a machine learning technique. How do these two areas fit together in your mind. Actually because the site focused teams all problems and can be served by machinery and when the approached this focusing problem with fines and attorney time service models are very important solutions to forecast team problems. Other side focused is very important in today's areas because you're low many many data so always Focused in problem. We find penser is very important and we also find iron man. Sarah map and armagh. Those are very classic. Run divided news time service motives and when we do couldn't this i remind saruman model. We're fans than actually the classical extre nation or classical. Modern for iran. serena is northern sufficient sarichichekli. Either way actually approach. I remember sarim from elisa angle which is spectral lenzi's digital delivery and in your system theory so we use elisa angle to do countries a romance. I remember motive defines onto loads dench a sufficient from the rich porno view. So this is the whole ground

Southwest Johnson University Carney Chung Penser Sarah Map Congress China Saruman Sarim Elisa Angle Lenzi Serena Iran Elisa