Margin

The term margin has many meanings:

• In telecommunications, margin has the following meanings:
  • In communications systems, the maximum degree of signal distortion that can be tolerated without affecting the restitution, i.e., without its being interpreted incorrectly by the decision circuit.
  • The allowable error rate, deviation from normal, or degradation of the performance of, a system or device.

Source: Federal Standard 1037C

• In publishing, the margin refers to the white space around text or illustrations in printed or displayed documents such as books.

• In business margins, margin can be the profit margin.

• In finance, a margin is the cash or collateral that a holder of a position in securities or in exchange traded derivatives is required to post to cover potential adverse movements in the value of the position.

• In manufacturing, gross margin is the difference between the energy, material and labor costs required to construct an object and the wholesale price obtained for it. This is not the same as profit, which must account for other costs such as depreciation, advertising, rental, engineering, etc. These and other expenses are paid for from gross margin, debt, or from start-up capitalization. Similar considerations apply to wholesale distribution or retail sales, and various service industries such as the exhibition of movies or the provision of gardening services.

• In Medicine, the tissue margins are the edges of an excision (or resection) surgical specimen. The margins are examined by a pathologist to evaluate the likelyhood that the disease (for example, a cancer) has been completely removed. "Clear margins," or "negative margins", means that the surgeon's cut was in healthy tissue, while "positive margins" means that the surgeon has cut through diseased tissue. This increases the likelyhood that some disease was left inside the patient and, therefore, that the disease will reccur.

• In machine learning, margin refers to the minimal distance of a sample to the hyperplane dividing positive and negative examples. Some machine learning methods, such as Support vector machine or Boosting, work by searching for a hyperplane that maximizes the margin.

• In execution of activity, "margin" of error means the 'edge' or 'point' at which one measures acceptable or unacceptable amounts or numbers of errors; the "cut-off" point.

The content of this page is retrieved from http://en.wikipedia.org/wiki/Margin under GFDL