Recuperación de estructuras latentes y estimación de parámetros en test multidimensionales ordinales: análisis comparado entre la TRI la TCT y el análisis factorial

A través de la realización de cuatro estudios basados en datos simulados y unidos por una preocupación común, aunque parcialmente independientes en su metodología y foco específico, esta tesis se plantea como objetivo general poner a prueba algunas de las ventajas que podría obtener un investigador aplicado al utilizar modelos clásicos de TRI frente a otras opciones, para de esta forma evaluar la posibilidad de obtener resultados relativamente equivalentes utilizando procedimientos alternativos. Específicamente, en esta tesis se pone a prueba: a) la calidad diferencial de las estimaciones de las propiedades de los ítems obtenidas por los procedimientos TRI frente a los factoriales; b) la validez de los procedimientos de escalamiento TRI frente a los de la TCT; y c) los límites de la propiedad de invarianza de las estimaciones de sujetos e ítems obtenidas con procedimientos TRI. La hipótesis general de esta tesis es que, en el tipo de condiciones en que se ha enfocado el estudio, las ventajas de utilizar modelos clásicos de TRI no serán demasiado amplias, con lo que existirán condiciones en las que emplear procedimientos alternativos será recomendable, dada su simplicidad relativa y su equivalencia de resultados. La primera investigación utiliza datos simulados para comparar las estimaciones de discriminación de ítems politómicos provenientes del AFI, frente a las posibles de obtener con distintas estrategias relacionadas con el análisis factorial clásico. La segunda investigación, también utiliza datos simulados para poner a prueba la equivalencia de las estimaciones de discriminación de los ítems obtenidas con los procedimientos TRI y AFI. La tercera investigación, utiliza dos estudios basados en datos simulados tanto para comparar los escalamientos de los sujetos obtenidos por la TRI y la TCT, como para aclarar las condiciones en que es posible esperar la presencia similitudes o diferencias entre ambos...

Two classes of metric spaces

The class of metric spaces (X,d) known as small-determined spaces, introduced by Garrido and Jaramillo, are properly defined by means of some type of real-valued Lipschitz functions on X. On the other hand, B-simple metric spaces introduced by Hejcman are defined in terms of some kind of bornologies of bounded subsets of X. In this note we present a common framework where both classes of metric spaces can be studied which allows us to see not only the relationships between them but also to obtain new internal characterizations of these metric properties.

Using WordNet relations and semantic classes in information retrieval tasks

In this paper we explore the use of semantic classes in an existing information retrieval system in order to improve its results. Thus, we use two different ontologies of semantic classes (WordNet domain and Basic Level Concepts) in order to re-rank the retrieved documents and obtain better recall and precision. Finally, we implement a new method for weighting the expanded terms taking into account the weights of the original query terms and their relations in WordNet with respect to the new ones (which have demonstrated to improve the results). The evaluation of these approaches was carried out in the CLEF Robust-WSD Task, obtaining an improvement of 1.8% in GMAP for the semantic classes approach and 10% in MAP employing the WordNet term weighting approach.

Ledger for the Classes of 1650-1663, 1649-1660

The leather-bound volume contains accounts for the Classes of 1650-1663 and Fellows. Many of the student bills were settled with payment in kind, often in the form of grain, apples, or livestock. The Steward's accounts with the College from 1656-1659 were entered on pages 295-303 and listed as "The steward is Creditor" and "The steward is Debitor." Credit entries include food, fuel, corn, and wages; debit entries list purchases and often identify the sellers.

Ledger for the Classes of 1703-1718, 1703-1731

The parchment-bound hardcover folio volume contains accounts for Fellows, Tutors, the President, and members of the Classes of 1703-1718. Payments were often made with bills of credit, money, and notes. The Steward's accounts with the College are interspersed in the volume.

Ledger for the Classes of 1723-1731, 1719-1755

The volume, bound in a modern leather binding, contains accounts for the Classes of 1723-1731. Leaves 108-118 list the Steward's account with Harvard College arranged monthly for the years 1721-1729.

Ledger for the Classes of 1732-1743, 1728-1749

The volume, bound in a modern green leather binding, contains accounts for the Classes of 1732-1743. The Steward's accounts with Harvard College are interspersed throughout the volume.

Ledger for the Classes of 1744-1749, 1733-1750

The volume, bound in a hardcover parchment binding, contains accounts for the Classes of 1744-1749.

Ledger for the Classes of 1795-1804, 1795-1804

The leather bound folio account book kept by Steward Caleb Gannett contains accounts for members of the Classes of 1795-1804, as well as accounts for President Joseph Willard, Tutors, Fellows, and Faculty. The volume also contains the Steward's accounts with the Harvard Treasurer, as well as his "Cash" accounting. The volume does not have the detailed transaction descriptions found in earlier ledgers; both Credit and Debit entries are described as being paid by, or being owed to, cash, balance, sundry accounts, or the Treasury.

Quarter bill book for the Classes of 1689-1723 and ledger, 1687-1720

Parchment hardcover bound volume arranged into two parts: Quarter bill book and the Steward's ledger. The Quarter bill book contains quarter bill tallies for the Classes of 1689-1723 arranged by seniority, and covers the bill period ending on October 21, 1687 through the period ending March 11, 1720.