Вы находитесь на странице: 1из 6

Information The manipulated and processed form of data is called information. It is more meaningful than data.

It is used for making decisions. Data is used as input for processing and information I output of this processing. Example Data collected from census is used to generate different type of information. The government can use it to determine the literacy rate in the country. Government can use the information in important decision to improve literacy rate.

3.1 Value of information The most comprehensive normative structure for information evaluation is the theory of information economics. This theory is based on economic and statistical decision theory and models the selection of optimal information in a cost/benet sense, by explicitly recognizing the decision maker's beliefs and preferences. A major concept in information economics is the value of 6information. This value is equivalent to the value of the information provided by the system and is the precise analogue for the more casual term "information usefulness".(Lawrence, 1999) In decision theory the demand price of information is dened as the maximum nonstochastic cost, payable from initial wealth prior to the receipt of any message, that makes DMs indierent between purchasing the information or not. For the cost-loss decision-making model, one way to measure the economic value of forecast probability estimators is in terms of the reduction in expected expense incurred by the decision maker. This measure involves a comparison of the expected expense using an estimator with that if climatological information alone were available to the decision maker. Information has a value sofar if it changes the decision maker's action, leading to a reduction in expected expense relative to the situation in which the information is not available. Both the value of perfect and imperfect information can be measured if the decision maker has the information. Let EC, EF and EP denote the expenses associated with climatological information, imperfect information and perfect information, respectively. The economic values of perfect VP and imperfect information VF , are given by VF = EC EF , VP = EC EP (2) The value of imperfect information satises 0 VF VP . In the further section, we show how VF behaves as a function of quality in the cost-loss ratio model. From the contingency table, for the prior information alone, a comparison of expected expenses (L if action a2 is taken and C if action a1 is taken) implies that action to protect should be taken if > C/L. 7In particular, the minimal expected expense is EC = L if C/L EC = C if C/L (3) A similar approach can be taken for the case of imperfect information. Taking an umbrella is optimal if P r [ = 1|Z] > C/L. Clearly, if p0 > C/L or

if p1 < C/L the optimal policy is identical to that of prior information. Therefore, the minimal expenses are the same. Only when p0 < C/L < p1 there will be a dierence in optimal policies. In this case, an umbrella should only be taken when rain is forecasted (Z = 1) and EF = (1 pz)p0L + pzC (4) where pz is dened as pz = ( p0)/(p1 p0) (5) Taking the dierence in the expected expenses (4-3) VF = (1 pz)(C p0L) if p0 C/L VF = ( p0)L pz(C p0L) if C/L p1 (6) 3.2 Quality of information The quality of a forecast has a broader scope to be examined under two main approaches, measure oriented and distribution oriented approaches (Murphy, 1993). Traditionally, forecast quality consisted of the computation of measures of the overall correspondence between forecasts and observations. The measure oriented approach tended to focus on one or two overall aspects of forecast quality, such as accuracy and skill. The perspective provided by the distribution oriented approach reveals that forecast quality is inherently multifaceted in nature. Correspondence between mean forecast and mean observation, overall strength of linear relationship between individual pairs of forecasts and observations, and average correspondence between individual pairs of forecasts and observations were some of the quality aspects. 8The expression (6) for the value of imperfect forecasts depends on two parameters, p0 and p1 related to the characteristics of the forecasts. To simplfy the the scientic quality of the imperfect weather forecast, long run relative frequency with which adverse weather is forecast is constrained to equal the long-run relative frequency of rain; that is P rZ = 1 = . This constraint implies that p0 = (1 p1)/(1 ) (7) The parameter p1, p1 1, completely determines imperfect information. Prior information and perfect information are special, limiting cases in which p1 = and p1 = 1 respectively. Linear transformation of p1 gives q = (p1 )/(1 ) (8) where q is a measure of relative distance between p1 and , making 0 q 1 with q = 0 for prior information and q = 1 for perfect information. q is a correlation coe-cient between the weather variable and the forecast of adverse weather Z. Hence, q is a relative measure of quality of imperfect information (forecast). It is important to note that paramaters p1 and q depend only on the probabilistic characteristics of the weather events and forecasts.

INFORMATION NEEDS

Within the field of user studies the investigation of 'information needs' has presented seemingly intractable problems. If we date user studies from 1948 and the Royal Society Scientific Information Conference, with its several surveys of users' information-seeking behaviour, [5] then the progress towards some theoretical understanding of the concept of 'information need' has been slow. This fact is recognized by virtually every commentator on the subject from Menzel [6] and Paisley [7] through the various authors in the ARIST volumes, [8] to Ford's review of 1977. [9] As well as drawing attention to this fact, the authors have tried to discover why it is so and have generally concluded that the reason lies in inadequate methodology and the failure to do research that is 'cumulative'. Attention has also been paid to the definitional problem of 'information need' [10] and the difficulty of separating the concept from 'wants', 'expressed demand', 'satisfied demand' and so on. However, while much of this work is very useful, the problem remains generally unresolved.

Figure 2: The context of information seeking

Partly, this is the result of a failure to identify the context within which information needs investigations are carried out. Figure 2 [11] is an attempt to show some of the possible contexts. (Figure 1 may be thought of as a sub-graph of Figure 2, centred on the user.) It is difficult in any two-dimensional diagram to convey the complexity of the 'real' world

and abstract elements of that real world. The 'universe of knowledge', for example, is an abstract concept which embraces all knowledgerelated objects, events and phenomena and, as such, clearly interacts with the 'physical universe'. To show the complex interactions of the physical and abstract universes, however, would involve a multidimensional diagram which would be extremely difficult, if not impossible, to express upon a sheet of paper. Accepting that difficulty, however, the 'user's life world' can be defined as the totality of experiences centred upon the individual as an information user. Within this life-world one important sub-world will be the world of work, within which will exist various 'reference groups' with which the user identifies: fellow professionals, the peer group within an organization and so on. The user will be in contact with a variety of 'information systems', only one of which is shown in the diagram, hence the indicated overlap with the user and his life-world. Within the information system two subsystems are shown: the 'mediator' (generally a living system, i.e. a human being) and the 'technology', used here in the general sense of whatever combination of techniques, tools and machines constitute the information-searching subsystem. The information system must have access to various 'embodiments of knowledge', phrased in this general way to indicate that such embodiments may be documents or living people. The lettered paths on the diagram are intended to show some of the possible search paths that may be used by the information seeker directly or used on his behalf by the information system and its subsystems.
Levels of information requirements There are three levels of information requirements for designing an MIS (Davis and Olson 1984). They are: At the organizational level, information requirements define an overall structure for the information system and specific applications and database. Application level requirements include social or behavioural - covering work organization objectives, individual roles and responsibility assumptions, and organizational policies - and technical, which are based on the information needed for the job to be performed. A significant part of the technical requirement is related

to outputs, inputs, stored data, structure and format of data and information processes. At the user level, database requirements can be classified as perceived by the user or as required for physical design of the database.