Mobile classifieds search
AssigneevMobo, Inc.
InventorsYan Or, Anand Kumar Sankaran, Madhu Gopinathan, Vinu Sundaresan
Life of this patent
Estimating the term…
Abstract
A classified ads server receives a search query from a mobile device operated by a user. The search query is parsed and a plurality of tokens is extracted from the search query. A query classification is determined based on the query, using a domain index, and query response candidates are selected from a content index based on the search query tokens and the query classification. A relevance score is generated for each of the query response candidates, and a classified ad is selected based on the relevance scores. The classified ad is sent to the user device.
Claims20
A computer-implemented method for providing access to classified ads via a mobile device, the method comprising a computer system performing the steps of:
- receiving a query from a mobile device, the query comprising one or more query tokens;
- normalizing the query tokens;
- based on the query tokens, selecting query response candidates from a content index for the classified ads;
- generating a relevance score for each query response candidate;
- selecting a classified ad based on the relevance scores for the query response candidates;
- and
- sending the selected classified ad to the mobile device;
- generating a term document matrix (TDM) based on the normalized query tokens and on tokens for the query response candidates;
- computing a term frequency-inverse document frequency (tf-idf) matrix based on the TDM, the tf-idf matrix co a vector for each query response candidate;
- normalizing the vectors of the tf-idf matrix;
- and
- computing a similarity score for each query response candidate based on the corresponding normalized vector of the tf-idf matrix.
A classified ads server system for providing access to classified ads via a mobile device, the server system configured for:
- receiving a query from a mobile device, the query comprising one or more query tokens;
- normalizing the query tokens;
- based on the query tokens, selecting query response candidates from a content index for the classified ads;
- generating a relevance score for each query response candidate;
- selecting a classified ad based on the relevance scores for the query response candidates;
- and
- sending the selected classified ad to the mobile device;
- generating a term document matrix (TDM) based on the normalized query tokens and on tokens for the query response candidates;
- computing a term frequency-inverse document frequency (tf-idf) matrix based on the TDM, the tf-idf matrix comprising a vector for each query response candidate;
- normalizing the vectors of the tf-idf matrix;
- and
- computing a similarity score for each query response candidate based on the corresponding normalized vector of the tf-idf matrix.
A non-transitory computer readable medium configured to
- store instructions executable by a computer processor, the instructions for:
- receiving a query from a mobile device, the query comprising one or more query tokens;
- normalizing the query tokens;
- based on the query tokens, selecting query response candidates from a content index for the classified ads;
- generating a relevance score for each query response candidate;
- selecting a classified ad based on the relevance scores for the query response candidates;
- and
- sending the selected classified ad to the mobile device;
- generating a term document matrix (TDM) based on the normalized query tokens and on tokens for the query response candidates;
- computing a term frequency-inverse document frequency (tf-idf) matrix based on the TDM, the tf-idf matrix comprising a vector for each query response candidate;
- normalizing the vectors of the tf-idf matrix;
- and
- computing a similarity score for each response candidate based on the corresponding normalized vector of the tf-idf matrix.
Dependencies are inferred from the claim text.