A core Course on Modeling Introduction to Modeling 0LAB0 0LBB0 0LCB0 0LDB0 S.26
Conclude phase: interpret result
Reflection: interpret result Right problem? (problem validation) Right concepts? (concepts validation) Right model? (model verification) Right outcome? (outcome verification) Right answer? (answer verification)
confidence assessible when modeled system modeled system model model modeling purpose modeling purpose are all known model modeled system purpose confidence needs represented by should fulfill with respect to When can we have confidence in a model?
example 1: elegant and simple model (elementary secondary school physics, say mechanics of levers and slides) elegant and simple model (elementary secondary school physics, say mechanics of levers and slides) modeled system: not explicitly defined modeled system: not explicitly defined purpose: to pass one’s exam purpose: to pass one’s exam When can we have confidence in a model?
example 1: elegant and simple model (elementary secondary school physics, say mechanics of levers and slides) elegant and simple model (elementary secondary school physics, say mechanics of levers and slides) modeled system: ship yard modeled system: ship yard purpose: to secure safe launch purpose: to secure safe launch When can we have confidence in a model?
example 1: elegant and simple model (elementary secondary school physics, say mechanics of levers and slides) elegant and simple model (elementary secondary school physics, say mechanics of levers and slides) modeled system: ship yard modeled system: ship yard purpose: to find direction of moving ship (uphill or downhill?) purpose: to find direction of moving ship (uphill or downhill?) When can we have confidence in a model?
example 2: model: full event log model: full event log modeled system: Internet traffic modeled system: Internet traffic purpose: diagnose performance bottlenecks purpose: diagnose performance bottlenecks When can we have confidence in a model?
example 2: model: full event log model: full event log modeled system: Internet traffic modeled system: Internet traffic purpose: document for archiving purpose: document for archiving jpg When can we have confidence in a model?
example 2: model: aggregated data model: aggregated data modeled system: Internet traffic modeled system: Internet traffic purpose: document for archiving purpose: document for archiving jpg When can we have confidence in a model?
example 2: model: aggregated data model: aggregated data modeled system: Internet traffic modeled system: Internet traffic purpose: analyse performance bottlenecks purpose: analyse performance bottlenecks When can we have confidence in a model?
Validation: is it the right model? consistency model - modeled system e.g. are cat.-III values correct? e.g. are cat.-III values correct? does the model behave intuitively? does the model behave intuitively? consistency model - purpose e.g. are cat.-II values conclusive? e.g. are cat.-II values conclusive? Confidence requires validation and verification
Validation: is it the right model? consistency model - modeled system e.g. are cat.-III values correct? e.g. are cat.-III values correct? does the model behave intuitively? does the model behave intuitively? consistency model - purpose e.g. are cat.-II values conclusive? e.g. are cat.-II values conclusive? Verification: is the model right? consistency conceptual - formal model e.g. are dimensions correct? e.g. are dimensions correct? is the graph a-cyclic? is the graph a-cyclic? are values within admitted bounds cf. types? are values within admitted bounds cf. types? Confidence requires validation and verification
Verification: is the model right? consistency conceptual - formal model e.g. are dimensions correct? e.g. are dimensions correct? is the graph a-cyclic? is the graph a-cyclic? are values within admitted bounds cf. types? are values within admitted bounds cf. types? Validation: is it the right model? consistency model - modeled system e.g. are cat.-III values correct? e.g. are cat.-III values correct? does the model behave intuitively? does the model behave intuitively? consistency model - purpose e.g. are cat.-II values conclusive? e.g. are cat.-II values conclusive? model modeled system purpose confidence needs represented by should fulfill with respect to conceptual & formal Confidence requires validation and verification
… based on Verificationaccuracyprecision Validation Confidence requires validation and verification
accuracyprecision
accuracyprecision high accuracylow precision high accuracy low precision low accuracy highprecision low accuracy high precision low accuracy low precision high accuracy high precision low bias large spreading low bias (offset, systematic error), large spreading low spreading large bias low spreading (noise, randomness), large bias large spreading, large bias low spreading, low bias a single result gives no information: look at ensembles Confidence requires validation and verification
accuracyprecision high accuracylow precision high accuracy low precision low accuracy highprecision low accuracy high precision low accuracy low precision high accuracy high precision Confidence requires validation and verification outlier (freak accident, miracle, …)
accuracyprecision high accuracylow precision high accuracy low precision low accuracy highprecision low accuracy high precision low accuracy low precision high accuracy high precision ?? ?? …accuracy can only be assessed with ground truth …assessment of precision needs no ground truth (reproducibility) Confidence requires validation and verification
Summary: we need model, modeled system and purpose to asses confidence validation: is it the right model? verification: is the model right? validation requires assessment of accuracy and precision