Decision Support and Business Intelligence Systems - Essay Example

Decision Support and Business Intelligence Systems al Affiliation Data warehouse According to Inmon (2002), data warehouse consists of integrated, subject oriented databases that bear the DSS function, at which each point the available data is relevant to a given time period. Mattison (2006), on the other hand, defines data warehouse as consisting of large group of computer initiatives whose major goal is to dig out information from legacy systems and make it usable to business persons, supporting their endeavor towards cost reduction and revenue advantage. Incorporation of data from a number of different sources into a repository data hub, storing not only current data, but also historical data to be used for developing trending reports so that the management can use the reports for comparisons, even if it the comparisons are annual, quarterly or even semi-annual. A detailed explanation of “subject oriented, integrated, time variant, non-updatable collection of data used in support of management decision-making processes” Subject oriented implies that data warehouse is utilized when examining specific subject field. For instance, in business, production can be the specific filed of examination. Integrated implies that a data warehouse incorporates data from a number of data source, for instance, from sales department and production department identify product A differently, however, the data warehouse will incorporate the two to identify the product in a single way. Time invariant implies that data, whether kept for 2 months, 7 months or even after a year is retrievable. Non updatable implies that once the data is entered into the database, it will not be mutilated and would not be changed, and, as a result, the historical data shall never be altered. What is natural language processing as it is related to text mining? Text mining refers to the unearthing and extraction of interesting, non trivial knowledge free from free and unstructured text (Kao & Poteet, 2007). The process encompasses information retrieval, and that is, document and website retrieval and even to classification and clustering of text. Natural language Processing, on the other hand, implies the endeavor to extract a complete meaning representation from such free text (Kao & Poteet, 2007). Natural Language processing has been developed to conform to a number of techniques that are naturally syntactically parsed utilizing information from proper grammar and lexicon, thereafter the ensuing information is then deduced semantically and thereafter applied to produce data. NLP can be profound, spanning every corner of the sentence, or even thin , straddling just few passages or phrases within sentences, and even utilize statistical ways to remove vagueness in word senses of the same sentence. What are some of the benefits and challenges of NLP? The unassuming challenge posed to the development of NLP is the rigidity of NLP unlike human speech which frequently is not precise as it is laden with ambiguity and the linguistic structure depends on complicated variables such as slang, dialects, and social settings. The benefit of NLP is often noticed in cloud computing and the development of service level management. In the medical profession, NLP helps in a mapping the large amounts of textual information to improving clinical outcomes. Laws are heavily dependent on semantics; machine translation using natural language processing has been applied in translation of all proceedings of the parliament of Canada and the European Union. Data mining, Text mining and Web mining This has been referred to as data or even knowledge discovery and is the process of data analysis from dissimilar angles and thereafter abridging into proper information Liu, B. (2007). Text mining refers to the unearthing and extraction of interesting, non trivial knowledge free from free and unstructured text. Miner (2012) refers to the process in principle as a simple exercise in communication. Web mining is the application of data mining processes to mechanically discover and extract relevant information from web documents and services. In general information discovered through web mining can be classified into three classes, namely; web activity, from server logs and web browser activity tracking, web graphs, and web content (Han & Kamber, 2006). Differences and Commonalities Text mining typically differs from web mining as the latter involves the user searching for what is already known and had already been written before hand. The tasks involve discarding all the entire irrelevant information to be left with the precise information one was searching for. As a result, the main stay is the inter-linkage of the information extracted to form consistent facts, which can be further proven through experimentation. The usage of texts in the processes marks the commonalities of the three, though the differences may appear the interaction is plainly on the usage of texts. Web mining, on the other hand is a sub set of data mining and involves the application of data mining processes to unearth patterns on the web (Han & Kamber, 2006). A distinction between text mining and data mining arises in the form of the patterns extracted as they are from natural languages instead of structured databases of facts. Web 2.0 revolution as it relates to BI There has not been a clear and concise definition of web 2.0, but the applicability of the tool to business intelligence has been noticed in the various filed the tool has been applied to and the results have been enormous (Solari, 2009).  For instance, if a business entity is to conduct a thorough competitive analysis, to determine important market trend that may have an impact on inventory turnover. The tool can be used to the in place BI solution to extract data on the competitive deals from customer management relationship and inventory force automation to reveal the customer gained and given away on the involvement of the rival enterprise. Further, the tool can merge the internal information together with information from analysts and a host of other relevant in house documents. The social networking sites that incorporate web 2.0 provide firms with practically all information on target market and their perceptions. Virtual world as it relates to BI The virtual world consists of different interactive environments and tools such as gaming consoles, internet amongst others. People constantly, through the interaction with the virtual world process large amounts of intelligence various forms such as; statistics, numbers, videos, sound bites, stock market tracking dash boards amongst others. This information is important in the business environment and aids in day to day decision making. McLennan (2008) quoted Bill Gates asserted that the virtual world provides the platform for collaboration, business intelligence and prioritization of the limited time that would prove difficult in the real world where processes move comparatively slower. How data can be divided into training and set test In the SQL 2012 server, one should separate the original data set at the stage of the mining structure. If one is using the explorer, the two can be divided into the following: Training set; load the full dataset, selection of the remove percentage to filter in the pre process panel, then selection of the appropriate percentage split, application of filter and finally saving the generated data to a new file. For the test set, the invertSelection property is set to true, the filter is applied and the generated new file is saved. Three steps of the ETL process ETL is the acronym for extraction, transformation and loading the three steps involves the extended meaning from extraction to loading. Step one involves extraction, and it involves connection to the source systems, and not only selection but also collection of the relevant data important for analytical processing, incorporating the data warehouse. The process converts data to a form suitable for the transformation process. Step two, transformation process, involves execution of a number of commands to the extracted data to convert it to the standard format. Common processes during transformation include clearing the duplicates and translation. Finally the loading process involves importing transformed and extracted data to the targeted warehouse or even database. The Balanced Scorecard The Balance scorecard is an analysis instrument developed to transform a company’s business strategy into identified, defined, quantifiable and specific objectives and to examine the performance of the company towards the achievement of the objectives. Biazzo & Garengo, (2012, 9) point out that the presence of a number of indicators coupled with the critical success factors, given the position of a company in the industry should be crucial and the defining statement of the balance scorecard. Yet, Kaplan & Norton (2013) suggest that balance scorecard is a break from factors not normally captured in the financial statements. Common myths associated with Data mining. `There are a number of myths associated with data mining, and often, people take that data mining is all about algorithms. The process of data mining is comprehensive and begins at the level of business goals mapping to evaluation and presentation of results. Secondly, it is often misconstrued to a process that is all about predictive accuracy. Third, for data mining to occur, then a data warehouse must be in place, often the latter is a prerequisite of the former. Fourth, data mining must be everything to do with a large number of data, however, a number of useful data mining processes can be worked on small data sets. Finally, data mining utilizes advanced technologies and of complex models and that majority of people not conversant with the IT things would not understand the workings. In contrary, data mining is not understood by the all and sundry within the IT industry, but an individual who has acquired a wealth of knowledge in the field. References Biazzo, S., & Garengo, P. (2012). Performance measurement with the balanced scorecard a practical approach to implementation within SMEs. Berlin, Springer. http://dx.doi.org/10.1007/978-3-642-24761-3. Han, J., & Kamber, M. (2006). Data mining: Concepts and techniques. Amsterdam: Elsevier. Inmon, W. H. (2002). Building the data warehouse. New York: J. Wiley. Kao, A., & Poteet, S. R. (2007). Natural language processing and text mining. New York: Springer. Kaplan, R. S., & Norton, D. P. (2013). The strategy-focused organization: how balanced scorecard companies thrive in the new business environment. Boston, Mass, Harvard Business School Press. Liu, B. (2007). Web data mining: Exploring hyperlinks, contents, and usage data. Berlin: Springer McLennan, K. J. (2008). The virtual world of work: How to gain competitive advantage through the virtual workplace. Charlotte, NC: Information Age Pub. Mattison, R. (2006). The data warehousing handbook. Oakwood Hills, Ill: XiT Press. Miner, G. (2012). Practical text mining and statistical analysis for non-structured text data applications. Waltham, MA: Academic Press. Solari, C. (2009). Security in a Web 2.0+ world: A standards based approach. Chichester, West Sussex: Wiley. Read More