Exam Decision Support System and Business Intelligence - Coursework Example

Exam Decision support system and business intelligence Lecturer: Question Data warehouse is a collection of data generated to support decision making because it is a repository of current and past data of interest to managers in an organization. Subject oriented data collection contains only information that is relevant in decision-making; moreover, a data warehouse being subject oriented allows users to establish the way a business is performing. Hence, data warehouse that is subject oriented provides a comprehensive view of an organization. The integrated aspect of data warehouses allow them to put data from various sources in consistent format and in accomplishing this, data warehouses first deal with naming conflicts and discrepancies in measuring units. Since data warehouses are presumed to be fully integrated, this helps in decision-making making processes. Time variant characteristic of data warehouses means they are capable of maintaining historical data. Therefore, data warehouses not only offer the prevailing status but also detect variations in relations when carrying out comparisons and forecasting, which are essential in decision-making. Time is one of the significant dimensions that are supported by all data warehouses, which enables the analysis of data from various sources in their multiple time points. One significant characteristic of data warehouses is that once data is entered in the system, the data cannot be changed or updated; however, data warehouses discard obsolete data by accepting new data. Essentially, all the above aspects enable data warehouses to be applications that provide decision support capability because it offers ready access to business information capable of creating business insight. Data warehouse as a repository of data provide processes that enable managers to make the necessary decision in business (Turban, Sharda, Delen & King, 2010). Question 2 Natural language processing as a component of text mining in artificial intelligence involves the study of problems associated with understanding natural human language with the aim of converting human language to formal representation easier for computer programs to manipulate. The aim of Natural language processing (NLP) is to shift from syntax-driven text manipulation to a true understanding and processing of natural language that values grammatical and semantic restraints within a given context. A prominent area where benefits of NLP have been witnessed is in customer relation management, which maximizes customer value through understanding better and effective ways of responding to actual needs as well as perceived needs of customers. Although, NLP has made significant progress in understanding natural human language, NLP still faces various challenges associated with its implementation. For instance, part-of-speech tagging is difficult because marking up terms in a speech with regard to nouns, verbs, adverbs, and adjectives relies on definition of the term and the context of its use. Moreover, text segmentation is a huge challenge because some languages like Chinese and Japanese lack single-word boundaries. In such cases, text parsing requires identification of word boundaries making it a difficult task. In addition, speech segmentation challenge arise when analyzing spoken language since sounds representing letters and words that follow each other are capable of blending. Another challenge in NLP is that words have more than one meaning; hence, to make sense of certain words necessitates consideration of the context. NLP also faces the challenge of syntax ambiguity because grammar in natural language is ambiguous which provides various possible sentence structures. Therefore, choosing the appropriate structure usually requires fusion of both semantic and contextual information. Moreover, NLP faces the challenges of regional accents and vocal impediments; as well, a speaker may consider a sentence an action while at the same time the sentence structure lacks information for defining the action (Turban, Sharda, Delen & King, 2010). Question 3 Text mining is a semi-automated process for mining patterns from huge quantities of amorphous data sources. Data mining is an approach used in recognizing valid, potentially useful, and ultimately comprehensible data patterns within structured databases. Web mining is a process for establishing intrinsic relations regarding useful information in web data that is expressed in terms of textual linkage or even usage information. Web mining is wide compared to text mining since the internet contains other components like multimedia and e-commerce data. However, although text mining as well as web mining remain diverse fields, most of the content within the internet is text-based. According to Kroeze, Matthee and Bothma (2007), close to 80% of the world’s online content is text based; hence, text mining forms an essential part in web mining. Text mining and data mining are similar they deal with automating the analysis of huge volumes of data and are useful in profiling an individual, companies and even groups of entities (Mena, 2003). Data mining primarily focuses on analyzing and discovering relations within structured data while text mining analyses specifically unstructured textual work in search of concepts and clusters in many documents or even web pages. The main difference between data mining and text mining is that text mining extensively uses lexical processing and analysis and other NLP techniques to highlight relations between words and clusters of documents based on their content. Web mining is similar to text mining because unstructured textual data in the form of web pages that are coded using HTML or XML and hyperlinks offer rich data for discovery and analysis (Turban, Sharda, Delen & King, 2010). Question 4 Web 2.0 revolution refers to the advances in web technology as well as applications such as blogs, RSS, mashups, user generated content, and social networks. Web 2.0 enhances creativity, sharing of information and collaboration, which differentiates it from the traditional web because of its improved collaboration between internet users, content providers, and enterprises. The business model that associates power with crowd emerges from web 2.0 making the model poses unlimited potential. The main aspect of web 2.0 is the worldwide spreading of innovative web sites and start-up firms and once the firms succeed other web sites appears across the globe. The main features provided by Web 2.0 involve its dynamic content, excellent experience for consumers, metadata, scalability, and freedom. The web techniques involved in development of creative applications in web 2.0 revolution are useful in making web pages more responsive. These techniques allow exchange of small data with the server without the need to reload the entire web page, which increases web page interactivity, loading speed and suability. Web 2.0 concepts resulted in web-based virtual communities together with hosting services like sites that support social networking and video sharing. Virtual worlds are simulated worlds developed by computer systems that allow the user to have an notion of being immersed. Virtual worlds are becoming essential means of reaching a wide consumer base and interacting with them in ways that were previously impossible. Companies that understand these applications and technologies and capitalize on them early in advance stand a chance of greatly improving internal business processes as well as marketing. The main advantages provided by the virtual worlds include improved collaboration with customers, suppliers, partners, and even internal users. Question 5 The core technical side of data warehousing process involves extraction, transformation, and load (ETL). ETL technologies are essential in processing and using data warehouses because the ETL process remains essential in any data-centric project. In ETL, extraction involves reading data from one or several databases while transformation involves converting the extracted data from its previous from to the needed form in order for the data to be kept in the data warehouse. The load aspect in ETL involves putting the data into the data warehouse (Turban, Sharda, Delen & King, 2010). Transformation takes place through rules or lookup tables by combining various sets of data. The three database functions are integrated into one tool that is capable of pulling data out of one or several databases and placing them in another consolidated database or rather a data warehouse. Therefore, ETL tools transport data from source to target while documenting the way data elements are transformed as they move from the source to the target. ETL processes are essential in data integration and data warehousing because the process loads integrated and cleansed data into the warehouse (Turban, Sharda, Delen & King, 2010). Question 6 Balanced scorecard (BSC) is widely acknowledged as a routine management method articulated by Kaplan and Norton. The system is useful in not only supplementing financial measures with nonfinancial measures but also communicating and implementing strategies. Hence, balanced scorecard is both a performance management methodology and a management methodology helpful in translating an organization’s financial, customer, internal process, and growth objective by targeting a set of actionable initiatives. As a measurement methodology, BSC overcomes the limitations of systems that mostly focus on finances because the methodology translates an organization’s vision into a set of connected financial and nonfinancial objectives, targets, and initiatives. As a strategic management methodology, BSc enable organizations to align actions with overall strategies through a series of interrelated six steps as established in the latest rendition by Norton and Kaplan (Turban, Sharda, Delen & King, 2010). References Kroeze J.K, Matthee M. C & Bothma T. J (2007). Differentiating between data-mining and text-mining Terminology. Retrieved 10 April 2014 from http://repository.up.ac.za/bitstream/handle/2263/3127/Kroeze_Differentiating%282004%29.pdf?sequence=1 Mena, J. (2003). Investigative data mining for security and criminal detection. Amtserdam [u.a.: Butterworth-Heinemann. Turban, E, Sharda R, Delen D & King D. (2010). Business intelligence: A managerial approach. Boston: Prentice Hall. Read More