International Business Machines Corporation

Nitaben A Patel

Kristin E McNeil

Dhruv A Bhatt

Bhargav Cheenepalli

Soomi Mun

A computer-implemented method, computer program product, and system for generating a targeted menu item recommendation are provided. The targeted menu item recommendation includes receiving a menu item recommendation request, generating search criteria for the menu item recommendation request, retrieving menu information regarding the search criteria, assigning weighted values to the retrieved information based on the text of the menu item information, preferences of the user, and social media association values, and generating the targeted menu item recommendation.

Determination of targeted food recommendation

Methods for text analysis of medical study data to extract predictive data. Natural language processing is performed on a document in a collection of documents to determine whether the document contains medical model data. In response to determining that the document contains medical model data, content relating to the medical model data in the document is annotated. A first medical model is generated based on the annotations for the identified medical model data and a certainty threshold In response to the certainty threshold meeting a user setting, the first medical model is added to a predictive model for determining a risk score, based on the analyzed data.

Determining a risk score using a predictive model and medical model data

Methods and apparatus, including computer program products, implementing and using techniques for text analysis of medical study data to extract predictive data. Natural language processing is performed on a document in a collection of documents to determine whether the document contains medical model data. In response to determining that the document contains medical model data, content relating to the medical model data in the document is annotated. A first medical model is generated based on the annotations for the identified medical model data and a certainty threshold In response to the certainty threshold meeting a user setting, the first medical model is added to a predictive model for determining a risk score, based on the analyzed data.

According to one embodiment of the present invention, a system for monitoring medical events for adverse effects includes at least one processor. The processor may be configured to identify one or more medical events within patient medical information by applying first analytics to the medical information. The medical information may include unstructured information. The processor is further configured to determine adverse effects of the identified medical events and corresponding time intervals of occurrence by applying second analytics to medical reference information, wherein effect information includes the identified medical events, determined adverse effects, and corresponding time intervals of occurrence. In addition, the processor is configured to determine from the effect information any adverse effects for a patient and the corresponding time intervals of occurrence. Finally, the processor is configured to notify a user of the determined adverse effects prior to initiation of the corresponding time intervals of occurrence.

Medical side effects tracking

Techniques presented herein describe filtering content rendered on a web browser. An analytics engine receives an aggregated feed to send to a requesting user. The aggregated feed includes one or more posts. For each of the posts, the analytics engine evaluates content of the post to determine whether a subject specified in a profile associated with the requesting user is present in the content. Upon determining that the subject is present in the content, the analytics removes the content from the post.

Techniques for filtering content presented in a web browser using content analytics

A method for identifying at least one contact associated with a user and an online communication network. The method may include receiving at least one search term or a social media post entered by the user. The method may also include performing text analytics on the received search term or the social media post to determine at least one subject. The method may further include identifying a plurality of contacts associated with the user, whereby the user and the plurality of contacts are associated with the online communication network. The method may additionally include performing text analytics on a plurality of communications associated with the identified plurality of contacts to determine a plurality of context information. The method may include comparing the determined subject to the determined plurality of context information. The method may include identifying at least one contact within the identified plurality of contacts based on the comparing.

Friend locator based on friend network and profile

A method, apparatus and article of manufacture for refining search results is disclosed. In one embodiment, the method comprises the steps of accepting a search query comprising one or more keywords, generating a first search result comprising one or more first search result members according to the search query, accepting a refinement command identifying one or more undesired first search result members, automatically refining the first search result to exclude the undesired first search result members and first search result members related to the undesired first search result members from the first search result, and providing the automatically refined search result.

System and method for refining search results

Provided are a method, computer program product, and system for relevant communications about unstructured information. Determining a relevant knowledge domain for the unstructured information. Generating one or more unstructured information annotations with a text analyzer, wherein the text analyzer is configured with at least one annotator related to the relevant knowledge domain. Generating one or more communication annotations for each of one or more communications with the configured text analyzer. Determining a relevance factor for each of the one or more communications based in part upon a comparison of the communication annotations to the information annotations. Presenting the communications with a relevance factor greater than a threshold relevance.

Relevancy of communications about unstructured information

Nutritional information of a recipe is gathered to determine a nutritional value table of a food recipe. A computing device may extract and analyze unstructured text of a food recipe to obtain a plurality of ingredients and a quantity of the plurality of ingredients. The computing device may access dietary preferences of a user. The nutritional information of the food recipe may be calculated using the nutritional value of each of the ingredients and complied into a nutritional value table. The recipe may be determined if the recipe corresponds with the dietary preferences of the user. If a recipe does not correspond with the dietary preferences, then an ingredient causes the recipe to not correspond with the dietary preferences is removed from the recipe creating an altered recipe. The nutritional value table of the altered recipe is displayed to the user.

Ingredient based nutritional information

An approach is provided in which a conversion manager receives a conversion request that identifies a conversion mode corresponding to a first category and a second category. The conversion manager identifies one or more first terms corresponding to the first category that are included in a page of text. As such, the conversion manager selects one or more second terms corresponding to the second category and, in turn, replaces the first terms with their corresponding second terms.

Growing community of inventors

Charlotte, NC, United States of America

Nitaben A Patel