Abstract Today’s environmental reports contain flaws in the acquired data. This master thesis has the mission to alleviate the estimations of those flaws. The data in question, originates from Swedish industrial facilities. The thesis involves data-treatment by statistical analysis, which is done through fitting a model by the means of analysis of variance and multilevel modeling. The thesis also involves gathering and work with data from databases, as well as systematic treatment, sorting, categorization and evaluation of the data material. Calculations are made through the SAS statistical analysis program, which rendered estimates of fixed, linear and random effects. The results are presented through graphs and numerical estimates in the later part of the report. Calculations for estimations of the grand pollutant totals are conducted. These are compared to the observed data for relevance. Alternative ways on working on the problem at hand is discussed, as well as problems that have appeared during the work on the master thesis. The relevant code and calculations are attached towards the end. / Sammanfattning Dagens miljörapportering har brister i den rapporterade datan. Examensarbetet har som avsikt att underlätta skattningen av den saknade datan i rapporteringen, vars data härstammar från svenska företagsutsläpp. Arbetet innebär en databehandling via statistisk analys, vilken utförs genom modellanpassning via variansanalys och flernivåmodellering. Arbetet omfattar även hämtning och bearbetning av datamaterial ifrån databaser, så väl som systematisk behandling, sortering, indelning och tolkning av dataobservationer. Beräkningar är utförda i SAS statistiska analysprogram, vilket renderat skattningar och representationer av termer till fasta, linjära och slumpartade effekter. Dessa presenteras med siffror och grafer i senare delen av rapporten. Skattning av totaler beräknas och jämförs med observerad data. Problem och alternativa angreppssätt diskuteras, samt kod och beräkningar bifogas.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:kth-99175 |
| Date | January 2012 |
| Creators | Berglund, David |
| Publisher | KTH, Matematik (Inst.) |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | Trita-MAT, 1401-2286 |
Page generated in 0.0025 seconds