Sampling methods: probability (simple random, stratified, cluster, multistage) vs non-probability (purposive, convenience). Simple random: equal independent chance, requires population list. Stratified: heterogeneous subpopulations. Cluster: cost-effective. Multistage: widely spread populations.
Forecasting: reliable estimates of uncertain future events using historical data. Types: demand, environmental, technological. Timing: short-range (<3 months), medium-range (3-12 months), long-range (3+ years). Methods: qualitative (personal opinion, Delphi), quantitative (freehand, smoothing, trend).
Sampling distribution: probability distribution of all possible sample statistic values. Mean sampling distribution properties: arithmetic mean equals population mean µ, standard deviation equals σ/√n. Central Limit Theorem: non-normal populations approximate normal (n≥30).
Sampling definition: selecting representative sample from population for analysis and inference. Needs: population movement, cost/time constraints, destructive testing. Bias types: undercoverage, non-response, wording. Sampling vs non-sampling errors. Population parameters (µ, σ²) vs sample statistics (X̄, s²).
Trend projection methods covering linear trend model, regression trend line equation (Ŷ = a + bx), normal equations for calculating a and b constants, properties of best fit trend line, solved examples for sugar factory and tourist arrivals data, seasonal effects measurement and seasonal index calculation (simple averages method, ratio-to-trend method, ratio-to-moving average method, link relatives method), deseasonalized values calculation, cyclical variations measurement, relative cyclical residual calculation with solved examples.