Description
This is a combined synopsis/solicitation for commercial items prepared in accordance with the format in Subpart 12.6, as supplemented with additional information included in this notice. This announcement constitutes the only solicitation; proposals are being requested and a written solicitation will not be issued. Solicitation number GED-RFQ-5235, this is a small business set-aside, NAICS 334611 Standard size 500. Item 1 – Decision Tree Software - The U.S. EPA, Office of Research and Development, National Health and Environmental Effect Research Laboratory, Gulf Ecology Division (GED) is currently seeking to purchase 2 named software licenses or 1 multi-user software license for developing tools (models) to assess and diagnose impairment in aquatic systems and the sources of associated stressors. Our research specifically addresses: 1) the development of methods and tools to enable states to monitor the condition of all waters and to predict the location of impaired waters; and 2) increasing our understanding of source/stressor and stressor/response relationships. Using principles of the Stressor Identification (SI) process supporting EPA’s guidance to assist the States in identifying causes of biological impairment, GED intends to develop the tools to diagnose the cause(s) of impairment from multiple stressors in coastal estuarine systems. Research includes the use of vast quantities of environmental data generated by GED-based projects, multiple EPA eco-divisions, and other federal and states agencies. GED scientists have determined that the best method for developing these models/tools will include the need to develop decision- or regression-trees for interpreting large, complex sets of environmental data. This type of software will allow us to thoroughly evaluate associations between varying types of data and pose well founded and comprehensive predictive models for determining the probability and source of estuarine resource impairment. Definitions: • All references to the term “software” from this point forward shall be defined as “classification and regression tree software”. Requirements: 1. The software shall use statistically sound and scientifically rigorous regression algorithms for generating decision models employing the methods described by Breiman, L., Friedman, J.H., Olshen, R. A., and Stone, C.J. 1983. Classification and Regression Trees. Wadsworth. 2. The software shall generate scientific and/or environmental data-driven regression models for building classification trees using target variable(s) that maybe: a. Continuous b. Discrete (categorical) c. Numeric or character d. Fractional 3. The software shall display, at a minimum, the following functions/features: a. Acquire data from ASCII-delimited formatted and database (including MSAccess and Oracle) resident sources. b. Provide model export options compatible with Microsoft Excel® and SAS® c. Maintain and report summary information about continuous statistics (mean, n, standard deviation, skewness, kurtosis, etc.). d. Maintain and report fully-weighed tabulation of distinct variables along with the quartile information and frequencies. e. Use entropy splitting rule relative to common levels f. Automatically use surrogate “splitting-variables” where values are missing (intelligent substitution) g. Generates the maximum number of node branches then optimizes the tree structure based on statistical importance. h. Determine the statistical importance of a variable and integrate the outcome across all of the tree splits. i. Use single variable splitting criteria - Gini, symmetric Gini, twoing, ordered twoing and class probability for classification trees. j. Use squares and least absolute deviation for regression trees. k. Execute automatic self validation procedures for model applicability (truthing) using relevant data from outside the development and training mode. l. Produce publication ready graphics. 4. The software shall be able to minimally accommodate a data matrix (array) containing a sample size of 25000 with ~ 200 variables. 5. The software must be able to provide a stable computing environment for manipulating large quantities of data without fatal application errors and using the standard, EPA-configured computing platform: a. Windows XP Profession Operating System Version 2002 Service Pack 1 b. Pentium® 4 2.80 GHz CPU; up to 2.0 GB RAM 6. The software vendor shall provide at least 1-year of technical support using the following modes: a. Email b. Phone 7. Award contingent on demonstration of key features outlined in this document. Line item 2 – Decision tree software – multi-user All items must be FOB destination. The following provisions apply: 52.212-1 - Instruction to Offerors – Commercial – 52.212-2 - Evaluation of Commercial items The evaluation for this procurement will be based on technical, price and past performance. 52.212-3 - Offeror Representations and Certifications – Commercial Items (offer must include a completed copy of this provision.) The following clause applies: 52.212-4 Contract Terms and Conditions – Commercial Items The closing date for this is procurement is September 7, 2005 at 10:00 a.m. EST For additional information: Bernice McCorvey, (850) 934-9225, Fax (850) 934-2400.