• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Progetto
 

Progetto

on

  • 343 views

 

Statistics

Views

Total Views
343
Views on SlideShare
343
Embed Views
0

Actions

Likes
0
Downloads
0
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Microsoft Word

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Progetto Progetto Document Transcript

    • Transition Facility Twinning Light Project Fiche Project Title Use of data mining for creation of analytical models in customs Project Number 2006/018-183-04-01-07 Twining Light No. Funding Programme 2006 Transition Facility, Unallocated Envelope Beneficiary Institution Customs Department under the Ministry of Finance Maximum Budget 159 200 EUR 1. Background and justification Under the strategy of Lithuanian Customs approved on April 2005 which is prepared following Council Resolution of 30 May 2001 on a strategy for the Customs Union, new priority tasks are foreseen for Customs bearing in mind its important role in protection of the society, therefore one of the long-term strategic goals of the Lithuanian Customs in the field of market, society protection and tax administration is to protect market and society from the damage caused by illegal international trafficking, to fight effectively against tax evasion, to improve tax accounting and collection, to preclude violation of customs legislation and other criminal activities. In order to achieve the goals stated above, the software designated for data mining has been obtained by Customs Criminal Service, under the Transition facility Project No. 2006/018-183-01-01 “Customs intelligence and statistic analysis”. . It is a powerful new technology tool with great potential to help customs focus on the most important information that is available in its data warehouses. With this data mining tool it possible to predict future trends and behaviors, allowing customs authorities to make proactive, knowledge-driven decisions. It scours databases for hidden patterns, finding predictive information that experts may miss because it lies outside their expectations. Most customs authorities already collect and refine massive quantities of data. Data mining techniques can be implemented rapidly on existing software and hardware platforms to enhance the value of existing information resources, and can be integrated with new products and systems as they are brought on-line. When implemented on high performance client/server or parallel processing computers, data mining tools can analyze massive databases to deliver answers to questions such as, "what are relations between raise of excise taxes and amount of seized cigarettes or price of cigarettes in black market and why?" At the moment Lithuanian Customs is implementing 2006 Transition Facility project No. 2006/018-183-01-01 “Customs intelligence and statistic analysis” and the project shall be completed at the beginning of 2009. Some activities of this project foresee trainings of the officers who use data mining software; however these trainings are oriented more on the use of the software and its possibilities instead of creating (programming) specific models according the customs needs. Data mining issues are only small part of the named project and in this case there is a need for very specific task – creation of the concrete model (or models) with help of available data mining software. Lithuanian Customs training centre also provides it’s own seminars on different aspects of investigation and analysis of customs fraud, however, these seminars mostly are organized by the local experts and are based more on general issues and are oriented for all customs officers, therefore there is a need for specific knowledge assistance only for officers who will work exclusively with data analysis i.e. data mining. 1
    • Lithuanian Customs administration still needs to improve its analytical capacities and this project would ensure more effective performance and use of obtained data mining software in the field of prevention of infringements of tax related legislation. Specifically, there is a need to create (to program) a specific model/-s for this software in order to solve a specific problems subject to customs law enforcement activities (e.g. to predict the trends of cigarette smuggling rate in relation to new legislation on excise tax). After procurement of the mentioned data mining tool it is clear that in order to achieve best results in using this software, assistance of more experienced EU experts is essential. Twinning partner’s input is necessary in order to receive an assistance of experienced officials dealing with data mining in their everyday duties, who could transfer their knowledge and expertise and help the Lithuanian Customs in the preparation of the specific models for using the mentioned software. In the future this system would empower the Lithuanian Customs to predict and analyze the trends of goods movements with possible gaps for infringements of customs laws, thus allowing prospective, proactive information delivery. It would also help to avoid losses in customs duties and ensure the proper flow of income to the EU budget. 2. Description of the Assignment 2.1. The beneficiaries The Beneficiary of this Twinning Light project is the Customs Department under the Ministry of Finance of the Republic of Lithuania. . The project will be implemented within the Customs Department, A. Jakšto str.1/25, 01105 Vilnius, Lithuania together with the Customs Criminal Service, Zalgirio str.127, LT-08217 Vilnius, Lithuania. The organisational structure of the Customs Department is as follows: Director General, four Deputy Directors General, the main divisions: Legal; Customs Legislation Harmonisation; Analysis of Statistics; Strategic Planning; Customs Procedures; Tariff; Tax Administration; Customs Work Management; Economic Entities Control; Internal Audit Service; Personnel and Training. The overall Lithuanian Customs organisation includes the Customs Department as headquarters, 5 regional customs administrations, and 41 Customs posts, Customs Criminal Service; Customs Information Systems Centre, Customs Training Centre and Customs Laboratory. Project Leader of the Beneficiary shall be Mr. Mantas Kausilas, Head of the Information analysis division of Customs Criminal Service: mantas.kausilas@cust.lt, tel.: (+370 5) 274 8033, fax: (+370 5) 274 80 24. The International Relations Division of Customs Department (Ms. Ana Burkovskiene, Chief Inspector of International Relations Division, ana.burkovskiene@cust.lt, tel.: (+370 5) 261 72 58, fax: (+370 5) 212 66 31) will be in charge of ensuring that the project would be implemented in accordance to Transition Facility rules. 2.2. Global and Specific Objectives The global objective of the project is to strengthen analytical capacities of the Lithuanian Customs in using the modern IT tools for law enforcement. The specific objective of the project is enhancement of the use of data mining for creation of specific models for the analytical solutions in the area of customs law enforcement. 2.3. Requested services 2
    • It is anticipated that the Twinning partner will assist the Lithuanian Customs counterparts in the following activities: • Analysis of the current situation regarding Data mining software used in the Lithuanian Customs, mainly based on interviews of the Lithuanian Customs officials and practical experience of the experts. Preparation of an analysis report including proposals on themes of analytical models for data mining tools used in the Lithuanian customs and their possible impact. • Preparation of at least two specific models for customs analysts using data mining software in the field of customs law enforcement while working together with Lithuanian customs officers. The Project Leader of the Twinning partner shall be responsible for the organisation of service delivery, reporting and other management issues. Indicative work plan: Input, man days Activities Expert Expert Total A B Project management 10 10 Analysing of current situation regarding 10 30 40 Data mining software used in the Lithuanian Customs and preparation of an analysis report Preparation of a specific models for 10 70 80 customs analysts using data mining software in the field of customs law enforcement TOTAL 30 100 130 2.4. Expected results: • An Analysis report, including proposed themes of specific models for data mining tools used in the Lithuanian customs. • At least two detailed functional models for data mining tools used in Lithuanian customs for analysts prepared. • Lithuanian customs officers working with data mining are trained and able to create their own models with available data mining tool. The Twinning partner will have to analyse the current situation regarding currently available data mining software used in the Lithuanian Customs and prepare an Analysis report including proposed themes for specific tasks, which should be created using available data with current data mining software. Lithuanian counterparts will provide all necessary technical tools and information in the form of documents available and interviews with Lithuanian customs officials. Preparation of specific models (working together with Lithuanian customs officers) with data mining software for the customs analysts will be a task for the Twinning partner, and Lithuanian counterparts will provide all necessary support. The Lithuanian customs officials will provide software and IT tools for the creation of models. 3
    • After the project it is foreseen that Lithuanian customs officers working with data mining tools will be able to create their own models according customs needs. The final versions of documents/ project results (agreed upon by both Lithuanian Customs and Twinning partner Project Leaders) shall be presented to the Steering Committee members for approval. 3. Expert profile General requirements for all experts: • Fluency in English. • Experience in conducting interviews in order to collect user requirements. Expert A – Project Leader: Qualifications and skills • Graduation of national customs academy and / or university or equivalent education in data mining (IT) and / or exact science and / or economics area. • Experience in project management. • Experience in working with projects funded by the European Union. General and specific professional experience • Experience in working in joint groups with other MS customs authorities in the customs field. • Experience in team work consisting of not less than 3 persons. • Experience in preparation of project implementation reports. • Not less than 5 years working experience in IT and analytical field. • Experience in the preparation of various technical documents of IT systems such as: functional and technical specifications, design documents etc. Experts B – Data mining expert: Qualifications and skills • Graduation of national customs academy and / or education in IT and / or exact science and / or economics area. • Knowledge of the EU Customs legislation and other documents governing the data mining processes in law enforcement. General and specific professional experience • Not less than 3 years working experience in a MS customs or other law enforcement administration. • Experience in the use of data mining tools in customs or other law enforcement (Tax administrations). • experience in the preparation of data mining models and work in law enforcement institution. • specific knowledge about data mining: time series analysis, retrospective data analysis, prospective data analysis, exploratory data analysis, artificial neural networks, clustering, genetic algorithms, linear regression, logistic regression, etc.. • Not less than 5 years working experience in design and/or development and/or implementation of the information systems. 4. Location and duration 4
    • The project will commence on 01/2009 and will end on 06/2009. The project duration will be of 6 months after the signature of the agreement. No Activities Input, 1 2 3 4 5 6 man days 1. Project management 10 2. Analysis of current situation regarding Data 40 mining software used in the in the Lithuanian Customs and preparation of an analysis report 3. Preparation of a specific models for customs 80 analysts using data mining software in the field of customs law enforcement TOTAL 130 The project shall be carried out in Vilnius, Lithuania. 5. Reporting and monitoring 5.1. Reporting requirements 5.1.1 This Twinning Light project is subject to the same monitoring procedures as standard Twinning. The Interim Quarterly Reports and a Final Report shall be prepared and submitted as defined in the Twinning Manual. Within 30 days from the commencement date of the contract the Twinning partner shall prepare an Inception Report including detailed work plan to achieve project objectives and expected results. Twinning partner shall submit the final report at the end of the assignment. The final report shall summarize and evaluate the results achieved, experiences and problems encountered and prepare recommendations for the further development. This report will describe the results of the project, compare these with the original objectives and assess the success of the project. It shall highlight any lessons learnt. The above-mentioned reports should be written in standard EU-Phare-format in English. The Twinning Project Leader shall submit signed copies of the aforementioned reports in three copies to the Customs Department. One copy of the approved reports shall be presented to the Central Project Management Agency. Working language of the project is English. 5.1.2. For coordination and monitoring of the project a steering committee will be established with the following members: representatives of the Customs Department, Customs Information Systems Centre, Customs Criminal Service etc Representatives from Ministry of Finance and Central Project Management Agency will participate in the meetings as observers. The role of the steering committee will be to review the project regularly, ensure that it is on schedule in all respects, and to take any major strategic decisions. The Steering Committee will meet on a quarterly basis (or more frequently if needed). 6. Total budget of the project 5
    • The maximum total budget for the project available is 159 200 Euro. 5000 EUR for translation and interpretation, 3000 EUR for audit, 3000 EUR for contingency costs. ANNEX TO PROJECT FICHE 1. Logical framework matrix in standard format. 2. Description of IT application used. 6
    • LOGFRAME PLANNING MATRIX LOGFRAME PLANNING MATRIX FOR Programme name and number Transition Facility 2006, TWL project: Use of data mining for creation of analytical models in customs Unallocated Envelope Contracting period expires Contract execution period expires 15/12/2008 15/12/2009 Total budget: 159 200 EUR TF budget: 159 200 EUR Overall objective Objectively verifiable indicators Sources of Verification • To strengthen analytical capacities of the • Number of seizures based on predicted Operational reports of the Lithuanian Customs in using the modern IT tools information is increased by 15 % till 2010. Lithuanian Customs for law enforcement • Amount of collected taxes by customs is increased up to 10 % till 2010. Project purpose Objectively verifiable indicators Sources of Verification Assumptions • Enhancement of the use of data mining for creation • At least two specific models created and used • Project final report • Experts qualification is of specific models for the analytical solutions in the with data mining tools currently available in sufficient. area of customs law enforcement Lithuanian customs Results Objectively verifiable indicators Sources of Verification Assumptions • An Analysis report, including proposed themes of • Analysis report is prepared and approved by • Project documents • Full commitment by specific models for data mining tools used in the the Steering Committee. • Project quarterly progress reports beneficiary ; Lithuanian customs. • At least two functional models are prepared • Operational reports of Customs • Prepared detailed functional models for data and approved by the Steering Committee. Criminal Service mining tools used in Lithuanian customs for • Lithuanian customs officers use created analysts. models in their daily work for submitting • Lithuanian customs officers working with data reports, forecasts, performing other analytical mining software are trained and able to create their activities own models with available data mining tool. 7
    • Activities Means Assumptions • Analysis of the current situation regarding Data Twinning Light contract • Work groups established by mining software used in the Lithuanian Customs, both parties mainly based on interviews of the Lithuanian • Experts qualification is Customs officials and practical experience of the sufficient experts. Preparation of an analysis report including proposals on themes of specific models for data mining tools used in the Lithuanian customs and their possible impact. • Preparation of specific models for customs analysts using data mining software in the field of customs law enforcement. Preconditions • Funding provided • Twinning partner selected in due time 8
    • Annex 2 to the Project Fiche IBM DB2 Data Warehouse Enterprise Edition v.9.1.2: - DWE Integrated Installer; - DB2 Enterprise Server Edition v9.1 UNIX; - DB2 Partition Feature v9.1; - IBM DB2 Performance Optimization Feature v9.1; - IBM DB2 Storage Optimization Feature v9.1; - DWE Administration Console; - DWE Cube Views; - DWE Design Studio; - DWE Intelligent Miner (Modeling, Scoring, Visualization); - DWE SQL Warehousing Tool; - WebSphere Application Server v6.0; - DB2 Alphablox v8.4; IBM DB2 Intelligent Miner for Data Intelligent Miner™ for Data Version 8.1 is an independent product that provides the following mining functions to build and apply mining models based on database or flat file data: • Associations mining function • Classification mining function including the following algorithms: o Neural Classification o Tree Classification • Clustering mining function including the following algorithms: o Distribution-based Clustering o Center-based Clustering • Regression mining functions including the following algorithms: o Neural Regression o Linear Regression o RBF Prediction • Processing functions The Processing functions can be used only on database tables. • Sequential Patterns mining function In DB2® Data Warehouse Edition, the Sequential Patterns mining function is called Sequence Rules mining function. • Similar Sequences mining function • Statistics functions 9
    • Intelligent Miner for Data Version 8.1 is a stand-alone data mining application (workbench) for end users and primarily statisticians with advanced data mining skills. It includes the Intelligent Miner Visualizers and the PMML conversion component of IM Scoring, which allows you to export mining models in PMML format. 10