Talk:Shock detector
 | This article is rated C-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | |||||||
| ||||||||
Interpretation edit
The section on Interpretation is acceptable but could be expanded. There could be several types of analyses for this type of situation. The analysis method listed in the citation ("Characterizing Shock Detector Performance", 1983, Packaging Technology) is old but still relevant. It is related to the packaging use of shock detectors but, in principle, it also relates to the ability of a shock detector to indicate probable personal injury.
Shock detectors can be activated by a shock exceeding the threshold of the indicator: it can be the peak g (acceleration level)of a shock pulse or by the energy content of the shock (impact velocity, drop severity). There is always some variation in the response of shock detectors to a given level of shock just as there is variation of response to a package or product being damaged (or the occurrence of injury). A statistical analysis of these responses can be useful to help understand the probabilities of correct vs false signals. The cited article investigates this.
Consider a hypothetical packaging situation where a shock detector on a packaged product is triggered 40% of the time at a certain severity of drop or impact: 60% of drops from that drop height would not activate the shock detector. Also consider a product which is damaged 30% of the time when subjected to the same severity of drop. These probabilities could be loaded into the contingency table as shown:
| Damage | No Damage | Total | |
|---|---|---|---|
| Detector Activated | 0.40 | ||
| Not Activated | 0.60 | ||
| Total | 0.30 | 0.70 | 1 |
The totals of the probabilities add up to one.
The four inner cells of the table can be calculated. For example, the probability of a correct signal of the shock detector being activated and the product being damaged is a joint probability: 0.40 x 0.30 = 0.12. The other inner cells are done similarly. Thus, knowing the performance of the shock detector and the sensitivity to shock of an item allows one to estimate the probabilities of correct signals, false positives, and false negatives.
| Damage | No Damage | Total | |
|---|---|---|---|
| Detector Activated | 0.12 | 0.28 | 0.40 |
| Not Activated | 0.18 | 0.42 | 0.60 |
| Total | 0.30 | 0.70 | 1 |
Of course, this analysis is only at one level of input shock. At a very low level of shock, the shock detector will not activate and there will not be damage. At a very severe level of shock, the shock detector will correctly indicate damage. Depending on the specific product fragility and the sensitivity of the shock detector, occurrences of false positives and negatives are usually when the input shock is close to the average levels of sensitivity for both the shock detector and the product.
The sensitivity of a product to damage (fragility) is usually known; should a shock detector be matched with that same fragility, higher fragility, or lower? When choosing a shock detector, one's tolerance to false signals must be considered. If no false negatives are allowed, a shock detector producing some false positives must be considered. Conversely, if false positives must be avoided, a shock detector with the possibility of false negatives needs to be accepted.
This discussion is offered for possible inclusion in the Wikipedia article; I am not making a specific proposal at this time. I will leave that up to other editors.
Merge with Shock indicator edit
Shock indicator is the same topic as shock detector. The article on Shock indicator should be merged into this article on Shock Detector. Pkgx (talk) 23:33, 9 September 2021 (UTC)
