Motakis, I. and Zaniolo, C. (1997). Temporal Aggregation in Active Database Rules. In ACM SIGMOD International Conference on Manage- ment of Data.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....T) atRisk(X, T 1) atRisk(X, T) atRisk (X, T 1) contacted (X, Y, T) atRisk (Y, T) This is Datalogls, the Templog formulation is similar. A query that is inductive on both time and data. New application area of Datalogls and extensions: operational semantics of active databases [Motakis and Zaniolo, 1997]. Motivation Marked Nulls and Constraints Queries (l;tcomplete temporal intormatlon Partial information: Sue stopped working for Microsoft and started to work for IBM before 1992 John worked for IBM before working for Microsoft Different granularities: The Beatles broke up in the ....

Motakis, I. and Zaniolo, C. (1997). Temporal Aggregation in Active Database Rules. In ACM SIGMOD International Conference on Manage- ment of Data.

....State Oriented Datalog Extensions. By extending Datalog with a notion of state, re)active production rules and deductive rules can be handled in a unified way, thereby combining the advantages of active and deductive rules. Two such (closely related) Datalog extensions are XY Datalog [Zan93,Zan95,MZ97] and Statelog [LHL95,LML96] see [KLS92] for an early precursor of the latter) The specification of operational aspects like composite event detection and coupling modes is possible in the logical language since the rules allow access to different database states even complex execution models ....

I. Motakis and C. Zaniolo. Temporal Aggregation in Active Database Rules. In ACM Intl. Conference on Management of Data (SIGMOD), pp. 440--451, Tucson, Arizona, 1997.

....between the computer vision algorithms and the database component, that is, essentially with a data modeling issue. The actual database structure goes beyond its scope. There is a rich literature in the theory of temporal databases, including the management of simultaneity and uncertainty [6, 9, 8]. Our database model has been described in [7, 10] 2 Overview The goal of the Presence technology is to collect events from an environment, store them, and allow users to search and view them, either live (i.e. while the actual events are happening) or from an archive of stored events. The ....

Iakovos Motakis and Carlo Zaniolo. Temporal aggregation in active database rules. In ACM SIGMOD International Conference on Management of Data, Tucson, AZ, USA, 13-15 May, pages 440--451, 1997. 16 Third Int. Workshop on Cooperative Distributed Vision

....simple but expressive language to specify policies. The design of the language has been strongly influenced by the action languages of Geffner and Bonet (Geffner Bonet 1998) and Gelfond and Lifschitz (Gelfond Lifschitz 1993) and the composite temporal event language of Motakis and Zaniolo (Motakis Zaniolo 1997). The semantics is founded on recent results on formal descriptions of action theories based on automata and their application to active databases. We summarize some complexity results on the hardness of evaluating polices and briefly describe the implementation of a policy server being used ....

....policies. The design of the language has been strongly influenced by the state language of actions of Geffner and Bonet (Geffner Bonet 1998) the action description language A of Gelfond and Lifschitz (Gelfond Lifschitz 1993) and the composite temporal event language of Motakis and Zaniolo (Motakis Zaniolo 1997). It uses the event condition action rule paradigm of active databases (Widom Ceri 1995) a successor of the production rule paradigm of languages such as OPS5 (Brownston et al. 1985) In fact, our language can be described as a real time specialized production rule system to define policies. ....

Motakis, I., and Zaniolo, C. 1997. Temporal aggregation in active database rules. In Proc. of SIGMOD.

....those for which there have been no sideeffects) and those for which external intervention may be required. They may also be viewed as those affecting the actions of the delayed transaction and those affecting other transactions. Some researchers have investigated the nature of timedependent rules [2, 8, 9, 10, 14, 15], while with the introduction of temporal databases, other researchers have investigated the accommodation of active rules in temporal databases [3, 4, 5, 6, 13, 16, 18] However, to date, no research has investigated whether the accommodation of active rules into temporal databases is able to ....

Motakis, I. and Zaniolo, C. Temporal aggregation in active databases rules. SIGMOD Rec., 26(2):440-451. 1997.

....framework for detecting action conflicts in policies and finding resolutions to these conflicts. The class of policies that we consider is limited to stateless transducers. We are currently extending our approach to deal with state. In particular, we accommodate sequence events (like in [20]) already a part of full PDL [15] This extension requires generalizing the semantics of policies and monitors to mappings from sequences of epochs to sequences of sets of actions. Also, we are studying monitors that are not based on cancelling conflicting actions but rather on delaying them ....

I. Motakis and C. Zaniolo. Temporal aggregation in active database rules. In Proc. of SIGMOD, Tucson, AZ, May 1997.

....and Toman [10] use PTL (past temporal logic) to express their temporal integrity constraints, Gertz and Lipeck [17] use FTL (future temporal logic) We reiterate that our goal is to be able to specify more than just integrity constraints. Recently Sistla and Wolfson [25] and Motakis and Zaniolo [21] use temporal and aggregate operators in their active database systems. But their use is towards having more expressive events and conditions (or a more expressive combination of both) in the triggers, not to declaratively specify the evolution of the database that is the goal or purpose of the ....

I. Motakis and C. Zaniolo. Temporal aggregation in active database rules. In SIGMOD 97, pages 440--451, 1997.

....both: ffl Datalog 1S and TempLog [Baudinet et al. 1993] BRICS Mini course on Temporal Databases This is Datalog1S , the Templog formulation is similar. A query that is inductive on both time and data. New application area of Datalog1S and extensions: operational semantics of active databases [Motakis and Zaniolo, 1997]. 92 2 Datalog 1S and Templog Find all the computers at risk where being at risk is defined in the following way: a computer is at risk at a given time if it has been earlier infected or it has been in contact with a computer already at risk. atRisk(X; T 1) infected(X; T ) atRisk(X; T ....

Motakis, I. and Zaniolo, C. (1997). Temporal Aggregation in Active Database Rules. In ACM SIGMOD International Conference on Management of Data.

....and implemented in SQL AG can depend on such order. Recent SQL extension for aggregates, supporting partition and windows for OLAP applications [25] also relies on the the order of data. Indeed in many applications, such as cumulative aggregates and moving windows aggregates used in timeseries [14], the fact that the data is sorted by their time stamps is part of the application logic. On the other hand, a direct application of online aggregates on stored data, normally relies on the fact that the data is not skewed [8] Thus, aggregates that are most useful in advanced applications are ....

.... 5] Col INT, Value INT, YorN INT) 6] TABLE mincol(Col INT) 7] TABLE summary(Col INT, Value INT, Yc INT, Nc INT, 8] KEY fCol,Valueg) 9] TABLE ginitable(Col INT, Gini INT) 10] 11] INITIALIZE : ITERATE : f [12] INSERT INTO treenodes [13] VALUES(RecId, iNode, iCol, iValue, iYorN) [14] UPDATE summary [15] SET Yc=Yc iYorN, Nc=Nc 1 iYorN [16] WHERE Col = iCol AND Value = iValue; 17] INSERT INTO summary [18] SELECT iCol, iValue, iYorN, 1 iYorN [19] WHERE SQLCODE=0; 20] 21] TERMINATE : f [22] INSERT INTO ginitable [23] SELECT Col, 24] ....

I. Motakis, C. Zaniolo, "Temporal Aggregation in Active Database Rules". In SIGMOD '97.

No context found.

I. Motakis and C. Zaniolo, "Temporal Aggregation in Active Database Rules." ACM SIGMOD, 1997.

.... SQL3 [20] 2 On the other hand, many important aggregate like computations cannot be express using the UDAs proposed in [20] For instance, time series analysis requires the computation of running sums or averages, and also of movingwindow aggregates (e.g. total sales for the last four weeks) [10]. The incremental computation of these aggregates is very simple given that the data is stored as a temporal sequence. For instance, the running total of sales is computed by updating the sum at each new sale and returning the accumulated result at the end of each day. These problems can be solved ....

I. Motakis, C. Zaniolo, "Temporal Aggregation in Active Database Rules". In SIGMOD'97.

....early returns discussed in the next section. It is important to observe that, while use mostly datamining examples for concreteness of discussion, the need for powerful UDAs is pervasive in advanced applications; for instance we have found them critical in temporal extensions to database languages [17]. 3 UDAs and Early Returns While the aggregate computations needed in a Bayesian classifier can be expressed using SQL built ins, this is not the case for most data mining algorithms. For instance the SPRINT classifier [23] chooses on which attribute and value to split next using a gini index: ....

....as in traditional aggregates. The computation of rollups, running aggregates, moving window aggregates, and many others becomes simple and efficient using the mechanism of early returns, which allows the generation of partial results while the computation of the aggregate is still in progress [17]. For instance, while final returns can be used to find a point of global minimun for a function, such as the gini function, early returns will be used to compute the points where local extrema occur (i.e. the valleys and the peaks) 2 4 Extended UDA and SQL3 At UCLA, we have developed the ....

I. Motakis, C. Zaniolo, "Temporal Aggregation in Active Database Rules". In SIGMOD'97.

....assumed until changed , i.e. until the default value. The event based implementation retains the advantages of the interval based implementation and has a closer relationship with composite event specification languages and provides an opportunity to combine active rules with temporal databases [MZ97]. 6 Conclusion We have shown that TDL is a natural temporal extension of Datalog. The same approach is also applicable to SQL and QBE. We used a point based data model at the conceptual level and took the point based reasoning approach beyond that of TSQL TP [Tom97, Tom98] by using aggregates to ....

I. Motakis and C. Zaniolo. Temporal Aggregation in Active Database Rules. In SIGMOD Record, Vol. 26, No. 2, pages 440-451, 1997

....ability of a computation to produce early returns is useful in many other situations besides on line aggregation, e.g. in time series analysis. Many different sorts of temporal aggregation, e.g. cumulative aggregation and moving window aggregation, are needed to perform time series analysis [MZ97]. For instance, a user might request the running sum of sales, i.e. the running sum of the sales from a given time. This computation is facilitated by the fact that the data is normally stored sorted by time, as in Table 2. Then, for each new sale, the running sum simply add the new amount to the ....

I. Motakis and C. Zaniolo, "Temporal Aggregation in Active Database Rules." ACM SIGMOD, 1997.

....help of SQL3 UDAs. While SQL3 UDAs could support the computation of the gini index, it cannot express many other useful UDAs. For instance, time series analysis requires the computation of running sums or averages, and also of moving window aggregates (e.g. total sales for the last four weeks) [5]. The incremental computation of these aggregates is very simple given that the data is stored as a temporal sequence. For instance, the running total of sales is computed by updating the sum at each new sale and returning the accumulated result at the end of each day. These problems can be solved ....

I. Motakis, C. Zaniolo, "Temporal Aggregation in Active Database Rules". In SIGMOD '97.

....minimum threshold. The ability of a computation to produce early returns is useful in many other situations besides online aggregation. For instance, to perform timeseries analysis we need many different sorts of temporal aggregation, such as cumulative aggregation, and moving window aggregation[21]. Thus, a user might request the running sum of sales, i.e. the running sum of the sales from a given time. This computation is facilitated by the fact that the data is normally stored sorted by time, as in Table 2. Then, for each new sale, the running sum simply adds the new amount to the ....

I. Motakis and C. Zaniolo, "Temporal Aggregation in Active Database Rules." ACM SIGMOD, 1997.

....this set of rules is not stratified, however, it is XY stratified. ffl Cumulative Context. In this context, for each component event, all occurrences of the event are accumulated until the composite event is detected. Then, aggregation can be performed on these component event occurrences. In [25], we show that temporal aggregation in active rules with composite can be easily incorporated within our semantics framework. 6.3. SAMOS and Petri Nets The most distinguishing feature of SAMOS is its event detection mechanism, which is based on colored Petri Nets [8] A colored Petri Net is an ....

....such as CLIPS can also provide a sound basis for implementation, due to their efficient rule activation mechanisms. The approach proposed in this paper lays the seeds for further research. We have already studied the issue of temporal aggregation in active rules and the results can be found in [25]. An important issue yet to be answered is the formal comparison of different composite event specification languages. In this respect, Datalog 1S provides a sound formal basis, due to the fact that its formal semantics is well understood and its expressive power w.r.t. to other languages ....

I. Motakis and C. Zaniolo. Temporal Aggregation in Active Database Rules. In Proceedings of the ACM SIGMOD Intl. Conf. on Management of Data, May 1997.

No context found.

I. Motakis, C. Zaniolo. "Temporal Aggregation in Active Database Rules," in ACM SIGMOD 1997

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC