Lesson 1 Flashcards
What is the goal of forecasting?
Forecasting aims to reduce uncertainty by estimating what will happen in the future.
What is qualitative Models?
Qualitative models are forecasting techniques based on judgment or subjective factor’s
What are the 3 forecasting techniques?
Qualitative models, Times Series Methods, Casual or regression based method.
What are forecasting models identified as?
Casual models if the variable to be forecast is influence be or correlated with other valuables included in the model.
What are time series models?
Time service model are forecasting technique that attempts to predict the future values of a variable by using only historical data on that one variable.
What is the process of time series?
The process involves extrapolation patterns and trends seen in past values onto the future values.
What are the components of a time series?
Is a sequence of values recorded at set interval of time. E.g hourly, daily, monthly, quarterly etc
These sequential values exhibit a variety of patterns that are broken down in four components.
As per render - 4 components are: As per Samuel level, trend, seasonality, noise (usually referred to as error-E
Trend
Seasonal
Cyclical
Random
What is the levels component?
Level describes the average value of the series.
It’s some what of a starting point - we can think of this as the Y-intercept
What is the trend component?
Trend is the general upward or down movement of the data
Trend may not be consistent in every time period due to random or other fluctuations
Trend is usually viewed over a longer period
What is the seasonal component?
Seasonality referees to a pattern of fluctuation above or blow some average level that repeats at regular intervals
What are cyclicals components?
The cyclical competent cover long-ten pattern that repeat ever several years
Cycle tend to be irregular in length and magnitude - difficult to predict
Think recession - economics activity recede and contract reaching down - later activity would pick up.
What is the random / error component?
Covers irregular, unpredictable variations in a time series - if the variable can not fit into trend, seasonal or cyclical variable then if is random
Any variation that don’t fit into t he previous components are considered noise
What are the 2 general forms of time-series models in statistics?
Multiplicative model - which assume demand is a product of the 4 component Demand = TSC*R
Additive model - which adds the component together - demand = T+S+C+R
What are the averaging techniques forecasting random variation?
Moving average
Weighted moving average
Exponential smoothing
What is a variable?
Is a characteristic or condition that changes or has different values for different individuals
