Filling Missing Climate Data Using Arithmetic mean method, Inverse Distance Weighting method MCMC

today we will see uh how to estimate uh the missing climate data using uh three methods the first one is automatic method and the second one is inversely starts with animated in the other the third method is multiple implication technique so estimations of missing precipitation data for this video tutorial involves three main steps the first step is selecting stations with fewer number of missing data the second step the main step is to arrange similar databases on day or on this month so the data to be estimated for missing uh should be the same should be within the same month or should be the same day the other uh is the other mistake is finally to estimate the missing climate data so we have a row station so for this demonstration video i have selected three main stations station a station b and station c as you can see station a is with no missing data but here station b with missing data there is n a around seven seven missing uh data are are within self months for station b efficiency is fully uh with no missing value as you can see the data is full and here station b there is with missing data so the first step is uh just the first step as you can see from from this uh sheet as you can see to select the stations with fewer number of missing data so we need to be take careful where we select a missing data so as well as as you can see as you can see this is no missing data but if i have to use a station a i need to check the data below 10 percent is missing uh else ignore the station so at 10 percent of the data from this data 10 percent uh below 10 percent should be missing or else i should rather ignore this station a so i i need to check uh if the missing is available in station a uh but here station b i have also i have to check the percentage here 10 missing value below 10 percent that is if it is above 10 percent the data is not good enough to do a data analysis so i need to ignore estimating for missing that so in this case the number is a seven there is seven as you can see seven so this equals two bracket seven divided by uh the numbers this is for 1981 so uh the sale individual share count will yield 365. so 365 minus without this without this missing value that means 365 minus seven so because bracket by nine times 100 so this value should be a less than 10 percent the final hitting turn so 1.95 percent which is less than 10 and this is okay so i can estimate this missing values because this the data of this station is missing data with fewer so i have to i can estimate because it's okay so the first step is to make sure to make sure to select a station with your numbers of missing data here i used three station station abc this is actually for demonstration purpose uh but i have mostly the better results comes from the the more station uh will yield the better approximation for missing data so it will be very nice to use a minimum of five station to predict a missing value so this is a fewer this is not unacceptable uh but this is a demonstration on how i can use uh these three methods so uh when you do by yourself you need to include the numbers of station to get the better value for missing data so the first one is arithmetic communicated so so don't forget to consider here the numbers of stitching as well as the numbers of missing data the other is to arrange similar databases one day or months so as you can see from stations in order to determine this missing data on safety number that is in order to calculate the missing values this i have to use these certain moments or else if i use self temperaments and general elements from the station for example from station a if i use general elements this is month which is zero value as you can see if i use this as uh to determine the missing for september months which is non-zero as you can see the data will not be good because uh we need to select september months of each station here september in september received a position also so i need to use this september month of each station so i will proceed to the ultimate method so as you can see pulsation nine months is september as you can see this is september month of 1981 here so i need to select the same environments as you can see this is for station a for station deforestation c in such a way you need to rearrange the data it's because it's a candlestick uh with similar data visitor dioramas then finally uh estimate the missing so but before going to estimation i need to say something about automatic animated yeah as you can see arithmetic method for a better estimation purpose data should be arranged with months day or year pattern so missing data is obtained by computing the ultimate coverage of the data corresponding to the nearest weather this is actually the equation to estimate use using this arithmetic method so where v naught is the estimated values of the missing data where v i is the values of the same parameter at the highest eyes nearest wizard station and n is the numbers of nearest station that is for for our case the numbers of miraculous station is tools this arithmetic method is mostly satisfactory if the gates are informally distributed over the area and the individual gauge measurement do not vary greatly about the mean according to a job so the second criteria needs to be fulfilled is normal and while precipitation of missing station should be below 10 percent of normal reforms of considered station i will proceed with the available station so we need to increase those index to get the better the best value for missing data so i will proceed how i can estimate this value equals to bracket this one plus this one close bracket divided by two here in there as you can see using value can be estimated in such a way so this is a better value if it is zero here zero here the value will be zero here as you can see how we can estimate using university source within it so for this method uh we need to have a location with xy complete data with station for every station you need to have xy data so the general equation can be this one this is the general allocation here for using this uh university translation method so uh how to calculate missing precipitation data using conversion starts with committed uh so the distance between two stations can be equated with the square root of x squared plus y squared and this inversely transmitted method is a better method it's because distance between station will be taken into account and mostly climate data especially varies with distance so this is a better method due to this reason so now uh i need to show you how i can or how we can estimate using this inverse distance with meter therefore the missing station is station b so we need to uh first prepare the distance between the missing station and the nearest wizard station so that is to calculate the distance between station a and station b and uh station between station b and station c uh so distance between station a and station b is this one so communicated into this one so the distance between station a a and b is equals to square root of a change in y that means h4 h4 minus h5 distance between this can be equally square root of x squared change in y square plus change next square with square roots then you get the distance between uh station a and b equals to this amount so the other is this is animator and the other distance between a station c with a missing station b so for this for this and such kind of problems distance from a missing station will still be first of all determined in such a way so from distance station c to station b can be created in such a way so for this case uh to differ between b and c here changing y square this one minus this one whole square plus this one minus this one double square then finally after saving both changes we need to squared it uh to get the distance between station b and c so it starts between station a and b one and this distance between p and c this one so if there is a station d you need to uh calculate distance between station d and station b so now let's proceed uh how we can use uh this general uh distance with eliminated equations so this one is equal to here as you can see the submissions of nearest weathering station divided by distance of weather station from the missing station so in this case we will break it so this one divided by distance from station b to station a that is this one close bracket so now plus because it's a summation now copy and break it the other is this one divided by distance from station c to b so this is it now close it the whole clause divided by the summation of one over distance i one divided by distance i that is this one close bracket so now class last open where one divided by this one so now close so hit enter so i need to here and this this distance are constant so only b8 and d8 will be con the first the other is a constant so function f4 insert eduard sign here hit enter now drag this is the final uh inverse distance weighted missing estimation numbers as you can see the other is multiple reputation technique i need to proceed to multiple implication techniques so we'll just use this method you need to activate the excel stats for me i i use excel stats you can use any other multiple uh implications into a program to estimate a missing data therefore here and to activate excel start versions of this excel yeah now this excel stats menu bar is appeared in such a way so quick preparing data here i'm missing data here yeah now as you can see i need to select the data range delete here so now i have to include these of each field so for september i need to use only save templar yeah which is safety number now it's okay click this one this is qualitative which is called the data so make sure this is quantitative data type the other resistance that you need to choose estimation the technique or method the other is remove draws with missing data or replaced by the mean the nearest neighbor nepal's for me i have to use this multiple implication microventing method so click this one if each field includes solutions of this data range includes the levels of each field so you need to activate this variable field click ok here there is an information 30 rows and three columns are selected including the estimations of this method so continue multiple imputation animated so this is the result so now i have to compare this copy make sure this copy so click here okay here value this is multiple multiple implication and the other is inverse distance inverse distance weighted committed the other is automatic limited needed so this one copy paste here without the database corresponding data here so now i need to see each plot so as you can see more or less inverse distance with limited uh is the red one and uh the arithmetic is almost uh similar hr type of the root here but the other multiple computation technique as you can see this almost green line uh very significantly with the other distribution uh limited but mostly uh this inverse distance within it is most preferable because those with data varies especially so this is all about today in such a way we can determine the missing climate that

Channel: Hydraulics and Hydrology Tutorials

Views: 8,858

Keywords: Missing Data Estimation, Inverse Distance Weighting Method, Arithmetic Mean Method, Multiple Imputation Technique, MCMC, XLSTAT, imputation, missing data, statistics

Id: _TU8uRK-0NI

Length: 17min 28sec (1048 seconds)

Published: Mon Nov 15 2021