Temporally-reweighted Chinese restaurant process mixtures for clustering, imputing, and forecasting multivariate time series. Saad, F. A. & Mansinghka, V. K. In *AISTATS 2018: Proceedings of the 21st International Conference on Artificial Intelligence and Statistics*, volume 84, of *Proceedings of Machine Learning Research*, pages 755–764, 2018. PMLR.

Paper Supplement Link Code abstract bibtex

Paper Supplement Link Code abstract bibtex

This article proposes a Bayesian nonparametric method for forecasting, imputation, and clustering in sparsely observed, multivariate time series data. The method is appropriate for jointly modeling hundreds of time series with widely varying, non-stationary dynamics. Given a collection of N time series, the Bayesian model first partitions them into independent clusters using a Chinese restaurant process prior. Within a cluster, all time series are modeled jointly using a novel “temporally-reweighted” extension of the Chinese restaurant process mixture. Markov chain Monte Carlo techniques are used to obtain samples from the posterior distribution, which are then used to form predictive inferences. We apply the technique to challenging forecasting and imputation tasks using seasonal flu data from the US Center for Disease Control and Prevention, demonstrating superior forecasting accuracy and competitive imputation accuracy as compared to multiple widely used baselines. We further show that the model discovers interpretable clusters in datasets with hundreds of time series, using macroeconomic data from the Gapminder Foundation.

@inproceedings{saad2018trcrpm, title = {Temporally-reweighted {Chinese} restaurant process mixtures for clustering, imputing, and forecasting multivariate time series}, author = {Saad, Feras A. and Mansinghka, Vikash K.}, booktitle = {AISTATS 2018: Proceedings of the 21st International Conference on Artificial Intelligence and Statistics}, series = {Proceedings of Machine Learning Research}, volume = 84, pages = {755--764}, publisher = {PMLR}, year = 2018, keywords = {probabilistic inference, multivariate time series, nonparametric Bayes, structure learning}, url_paper = {http://proceedings.mlr.press/v84/saad18a/saad18a.pdf}, url_supplement = {http://proceedings.mlr.press/v84/saad18a/saad18a-supp.pdf}, url_link = {http://proceedings.mlr.press/v84/saad18a.html}, url_code = {https://github.com/probcomp/trcrpm}, abstract = {This article proposes a Bayesian nonparametric method for forecasting, imputation, and clustering in sparsely observed, multivariate time series data. The method is appropriate for jointly modeling hundreds of time series with widely varying, non-stationary dynamics. Given a collection of N time series, the Bayesian model first partitions them into independent clusters using a Chinese restaurant process prior. Within a cluster, all time series are modeled jointly using a novel “temporally-reweighted” extension of the Chinese restaurant process mixture. Markov chain Monte Carlo techniques are used to obtain samples from the posterior distribution, which are then used to form predictive inferences. We apply the technique to challenging forecasting and imputation tasks using seasonal flu data from the US Center for Disease Control and Prevention, demonstrating superior forecasting accuracy and competitive imputation accuracy as compared to multiple widely used baselines. We further show that the model discovers interpretable clusters in datasets with hundreds of time series, using macroeconomic data from the Gapminder Foundation.}, }

Downloads: 0

{"_id":"Tp4DBiu32tXdeG9CY","bibbaseid":"saad-mansinghka-temporallyreweightedchineserestaurantprocessmixturesforclusteringimputingandforecastingmultivariatetimeseries-2018","downloads":19,"creationDate":"2018-04-20T17:49:55.782Z","title":"Temporally-reweighted Chinese restaurant process mixtures for clustering, imputing, and forecasting multivariate time series","author_short":["Saad, F. A.","Mansinghka, V. K."],"year":2018,"bibtype":"inproceedings","biburl":"probcomp.csail.mit.edu/papers.bib","bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"Temporally-reweighted Chinese restaurant process mixtures for clustering, imputing, and forecasting multivariate time series","author":[{"propositions":[],"lastnames":["Saad"],"firstnames":["Feras","A."],"suffixes":[]},{"propositions":[],"lastnames":["Mansinghka"],"firstnames":["Vikash","K."],"suffixes":[]}],"booktitle":"AISTATS 2018: Proceedings of the 21st International Conference on Artificial Intelligence and Statistics","series":"Proceedings of Machine Learning Research","volume":"84","pages":"755–764","publisher":"PMLR","year":"2018","keywords":"probabilistic inference, multivariate time series, nonparametric Bayes, structure learning","url_paper":"http://proceedings.mlr.press/v84/saad18a/saad18a.pdf","url_supplement":"http://proceedings.mlr.press/v84/saad18a/saad18a-supp.pdf","url_link":"http://proceedings.mlr.press/v84/saad18a.html","url_code":"https://github.com/probcomp/trcrpm","abstract":"This article proposes a Bayesian nonparametric method for forecasting, imputation, and clustering in sparsely observed, multivariate time series data. The method is appropriate for jointly modeling hundreds of time series with widely varying, non-stationary dynamics. Given a collection of N time series, the Bayesian model first partitions them into independent clusters using a Chinese restaurant process prior. Within a cluster, all time series are modeled jointly using a novel “temporally-reweighted” extension of the Chinese restaurant process mixture. Markov chain Monte Carlo techniques are used to obtain samples from the posterior distribution, which are then used to form predictive inferences. We apply the technique to challenging forecasting and imputation tasks using seasonal flu data from the US Center for Disease Control and Prevention, demonstrating superior forecasting accuracy and competitive imputation accuracy as compared to multiple widely used baselines. We further show that the model discovers interpretable clusters in datasets with hundreds of time series, using macroeconomic data from the Gapminder Foundation.","bibtex":"@inproceedings{saad2018trcrpm,\ntitle = {Temporally-reweighted {Chinese} restaurant process mixtures for clustering, imputing, and forecasting multivariate time series},\nauthor = {Saad, Feras A. and Mansinghka, Vikash K.},\nbooktitle = {AISTATS 2018: Proceedings of the 21st International Conference on Artificial Intelligence and Statistics},\nseries = {Proceedings of Machine Learning Research},\nvolume = 84,\npages = {755--764},\npublisher = {PMLR},\nyear = 2018,\nkeywords = {probabilistic inference, multivariate time series, nonparametric Bayes, structure learning},\nurl_paper = {http://proceedings.mlr.press/v84/saad18a/saad18a.pdf},\nurl_supplement = {http://proceedings.mlr.press/v84/saad18a/saad18a-supp.pdf},\nurl_link = {http://proceedings.mlr.press/v84/saad18a.html},\nurl_code = {https://github.com/probcomp/trcrpm},\nabstract = {This article proposes a Bayesian nonparametric method for forecasting, imputation, and clustering in sparsely observed, multivariate time series data. The method is appropriate for jointly modeling hundreds of time series with widely varying, non-stationary dynamics. Given a collection of N time series, the Bayesian model first partitions them into independent clusters using a Chinese restaurant process prior. Within a cluster, all time series are modeled jointly using a novel “temporally-reweighted” extension of the Chinese restaurant process mixture. Markov chain Monte Carlo techniques are used to obtain samples from the posterior distribution, which are then used to form predictive inferences. We apply the technique to challenging forecasting and imputation tasks using seasonal flu data from the US Center for Disease Control and Prevention, demonstrating superior forecasting accuracy and competitive imputation accuracy as compared to multiple widely used baselines. We further show that the model discovers interpretable clusters in datasets with hundreds of time series, using macroeconomic data from the Gapminder Foundation.},\n}\n\n","author_short":["Saad, F. A.","Mansinghka, V. K."],"key":"saad2018trcrpm","id":"saad2018trcrpm","bibbaseid":"saad-mansinghka-temporallyreweightedchineserestaurantprocessmixturesforclusteringimputingandforecastingmultivariatetimeseries-2018","role":"author","urls":{" paper":"http://proceedings.mlr.press/v84/saad18a/saad18a.pdf"," supplement":"http://proceedings.mlr.press/v84/saad18a/saad18a-supp.pdf"," link":"http://proceedings.mlr.press/v84/saad18a.html"," code":"https://github.com/probcomp/trcrpm"},"keyword":["probabilistic inference","multivariate time series","nonparametric Bayes","structure learning"],"metadata":{"authorlinks":{"mansinghka, v":"http://probcomp.csail.mit.edu/","saad, f":"http://fsaad.mit.edu/"}}},"search_terms":["temporally","reweighted","chinese","restaurant","process","mixtures","clustering","imputing","forecasting","multivariate","time","series","saad","mansinghka"],"keywords":["probabilistic inference","multivariate time series","nonparametric bayes","structure learning"],"authorIDs":["24ZZQYHs34DAa9Kr6","28Pw99RLurcmHxnQi","2Hocqzo62AGpazysG","2Rnuo8am6ZJ3zaiWF","2xQ3QSvHeSkx37vFX","3BTqPkyp87F3LdfbT","3DwpmEKdz76sHpmms","3aLDT4CEdr9WQpBic","3h44YL555EzhtTyov","3sWyAGovNzzFZeFzb","3zMd5SAGagiqnQFLx","464Mzt2HES7jktmMC","4QwQPXfYWxAfrqD3x","4YcH7NwtDKS68uEYh","4rod99bRJ9kDkjxjq","4uKY4xENBpZT6dDoq","585Ntjij9LqzHgJec","58e5112f03d5065b4a0000a9","5GyX8nSxnzPXAAriL","5JatimquQYofrgtYy","5XYQacne4xXLAyusX","5bcf251ecd6cac10000000a8","5de6d7c1abd988de01000108","5de6dd20abd988de01000182","5de81cfb4e3c5af301000065","5de823cd4e3c5af3010000fa","5de83b1d8cf0fbde01000084","5de83d928cf0fbde010000ab","5de8698c8ff138de01000147","5de88cbb2c5eb2df0100002e","5de8a8213cfb74df01000035","5de8f0e8978afbdf01000156","5de91f095d589edf010000c7","5de96d2e768c31df0100005a","5de96e66768c31df0100007a","5deaa35d03c11ade01000047","5deab8db03c11ade0100026d","5deab9dd03c11ade01000286","5deac944cd660bde0100014b","5ded4f29a49d79de010000de","5ded89c85fefb3de010000ae","5def414066cfe0df01000046","5defe8db14db5cdf01000055","5deff4d314db5cdf0100012e","5df06853e4ce32df0100011c","5df172f857cfb4de0100003f","5df1cf751070c8ef010000ac","5df1d5751070c8ef010000fc","5df264cd63aac8df01000188","5df26ca127cd2fde01000015","5df2872327cd2fde010001d7","5df2a9960f18f1df01000071","5df2c3bc79c00ade01000032","5df2c9ac79c00ade01000083","5df2e966b91ab0de010000b9","5df358afd617e5de010000b1","5df3ad7cec6029de010000e0","5df3ba9bec6029de010001ff","5df3c3db580920de01000073","5df41f6d9d2522de010000b6","5df4b300a50a84df010000d5","5df6ad3a31a37ade0100013e","5df6ca73f9318fdf01000177","5df703a9df3bb9f20100016a","5df703ecdf3bb9f201000171","5df76ca5952c76df0100011e","5df7b1c892a8e4df01000005","5df7bbc092a8e4df0100008a","5df7e63edc100cde010001c1","5df86547b99bcff30100005e","5df8c1cae6b510df01000118","5df92329d04b27df0100004c","5df92961d04b27df010000b8","5dfa9539669fc3df01000090","5dfb3d1ee04f92df0100005e","5dfbb80bf6f0aede010000ec","5dfbd8534705b7de01000164","5dfbf72d4d4fb9de01000196","5dfbff4fb371afde0100004b","5dff6a86cb2e2bde0100005a","5e038ae0d6cccbdf010000c0","5e0429e6c51eb3de01000070","5e049d4410c665de010000ab","5e04a0f310c665de010000db","5e053342709177de01000196","5e05ca85459a30df01000068","5e068dfcd4589cdf01000061","5e095d9572022bde010000af","5e0b6a25e73cd6de0100009a","5e0df6ae45d2fdde010000c4","5e0f3a572c4a31df01000054","5e0feb62063b5cde0100007e","5e10ecb745c12cde010000b5","5e11ba8b2af340df010000ab","5e12af1b3f181ade01000090","5e1352de697554de010000ea","5e135a70697554de0100016f","5e13adfa0d0b99de01000133","5e14863808c265de010000d3","5e159116f1f31adf01000135","5e15aa3ee8bfcff20100016d","5e162fc6df1bb4de01000184","5e166c3a505a61df010008b7","5e16ee965e8bcfdf010000be","5e175ca74df69dde010000a9","5e179346d954d8df01000043","5e18969f779abfdf010000ef","5e18a2b8779abfdf010001ee","5e192f00eaca22df01000058","5e1a11a2cde53bde01000116","5e1a2e059fbdddde0100010d","5e1aed275f3d2cdf01000123","5e1b379c749a45df01000050","5e1b8a407c0fe1df01000077","5e1ba2c361cb16df01000032","5e1c5707e556c6de01000100","5e1c5c86e556c6de01000176","5e1cdc1dabed9bde010000bf","5e1dde194986afdf0100015b","5e1e26c2407a20de01000019","5e1e523e2e41a7de010000fb","5e1e5c672e41a7de01000166","5e1f240307379ade0100001d","5e1f2d8107379ade010000b6","5e1f7418e8f5ddde0100022b","5e1f7f8008195af301000090","5e20acde5c2065de0100000d","5e20b7445c2065de010000ac","5e20ba3d5c2065de010000db","5e210a69c63e88df01000071","5e219f263ef35cdf010000c7","5e248ed68c3885df0100007a","5e24b9fb1a6264de0100011c","5e251a51b55b9edf01000074","5e25cf83f299d4de010001cc","5e2639f624c8a6de010000cb","5e2666dd581147f201000031","5e270bf4f51e02de01000009","5e270ca3f51e02de0100001f","5e2778b858994fde01000224","5e28a71d88416fde01000173","5e28b1ca6acacbdf01000016","5e28d496a3df5bdf01000064","5e28e19ca3df5bdf010000ff","5e2b545e6366e2df01000002","5e2c8658ce5606de01000102","5e2ca62ecca05fde01000003","5e2cbca6cca05fde01000118","5e2de707133023de0100009b","5e2e2f1ecc9e90e401000118","5e2ea7efb84405df0100007c","5e2ec219c2015cde01000055","5e2ef9e4e374eede0100006c","5e2effa8e374eede01000104","5e2f5fa326e5cadf01000312","5e30a8ccc99510de0100010b","5e3136a55a3ceade010000be","5e3137a05a3ceade010000d1","5e31c760b5384fde01000176","5e31c93cb5384fde01000193","5e31c98ab5384fde0100019b","5e31cd5bb5384fde0100023d","5e31ce27b5384fde01000269","5e31d2cabba4d0de01000061","5e31d5f6bba4d0de01000095","5e329a80e17accde01000069","5e3433c641f782de01000167","5e3495dc53b794de010000b4","5e34be67ed2109df010000b7","5e353ab3ba591ddf0100000f","5e3609e5591ba4de010000b1","5e36a8d77b975dde01000016","5e36dd76b26a0fde01000001","5e36e37db26a0fde0100003c","5e37827bac7fe8de0100014e","5e382c240691b8de010001ee","5e383748ca2b4bde010000d9","5e383b96ca2b4bde0100014f","5e384061ca2b4bde010001b8","5e38414bca2b4bde010001e1","5e386555ccda85de01000635","5e386a131f8af9e001000040","5e3880661f8af9e0010001d9","5e3881a71f8af9e0010001ee","5e388c7d030bcadf010000a3","5e397ce6346d7cde010001fd","5e3991c6d14579de010001db","5e39bfbfd6538cdf0100004b","5e39cc6dd6538cdf0100015e","5e39d34cd6538cdf010001d1","5e39f362aa2adade01000038","5e3ad4901b85fadf01000033","5e3ae1bc1b85fadf01000133","5e3b098555f0f2df010001b9","5e3b2960255dcade01000047","5e3b3442255dcade0100011b","5e3c2ec634cd37de010000e9","5e3c6e9d67788ede01000110","5e3c9dcdf99721de01000076","5e3c9fbbf99721de010000a0","5e3cdde75cd237de010000a4","5e3d2557c405ecde01000034","5e3d7fc696e576de010000c4","5e3ecf0086a596de01000058","5e3f47c777baf5df0100009f","5e3f9974cecf86de0100017a","5e40136f17f17dde0100019b","5e4041f7668183de0100008a","5e41b6a30b4861de0100017a","5e41ef93a9454ade010001b0","5e4202a4ebe241de010000d7","5e420517ebe241de010000ff","5e428bc75ea111df010000bd","5e42c7eee7fe39df010000e7","5e436014a37866de01000174","5e43f473547b4cdf0100004c","5e4418e1fdc393de010000de","5e44454edf3c2af30100000a","5e445a86df3c2af301000191","5e4487f5ec14b3de01000094","5e44ce167759a7df0100013e","5e44ec83ab9cedde0100011d","5e453154605639de01000088","5e46bd088573d1de010000e5","5e49612b16841dde01000076","5e498800407e6cde01000103","5e49c95eb63120f201000048","5e4abc2c15f6c7df010001ef","5e4b94db2d0ef5de010000aa","5e4bad3b733a18de010000c2","5e4de9e752c311f201000154","5e4e0047cc196bde0100014c","5e4ea27764b624de010000a8","5e4ea58d64b624de010000e0","5e4f42208a3535f301000013","5e50afb4e3d144df01000002","5e50b11be3d144df01000013","5e51f6d68240c0df01000079","5e53fa64d26e87df010001f0","5e54502988d190df01000016","5e553b5dca58a8df010000e3","5e55504be89e5fde01000037","5e557e89e11ab9df0100010f","5e55aa627d0846de01000096","5e55f822819fabdf01000028","5e56c16a96127bde010000ac","5e57501b18f14bdf01000058","5e580681a38020de010001c7","5e59268fe60e02de01000035","5e5c1fcb15d8f5de01000029","5e5c3c7e68f281de01000030","5e5cada2a9598ddf01000029","5e5d3b9b73eb2edf0100008b","5e5d5e1ead47bcde010000d8","5e5d676cad47bcde01000156","5e5fb8c519c3fade01000205","5e60088613e3aede01000156","5e6008bb13e3aede01000165","5e601354c064fcde01000071","5e601946c064fcde010000ed","5e6019aec064fcde01000101","5e601a2ec064fcde0100010d","5e601a41c064fcde01000111","5e601a54c064fcde01000117","5e601a57c064fcde01000119","5e601a69c064fcde01000120","5e601a86c064fcde01000128","5e601afec064fcde01000156","5e601b52c064fcde01000173","5e601caec064fcde010001d8","5e601d32c064fcde010001e4","5e601d58c064fcde010001eb","5e601db8c064fcde010001fe","5e601e01c064fcde01000209","5e601e1fc064fcde0100020c","5e6035b2182590df01000070","5e6108ef31c7d3de010001ef","5e613e4697c182e901000170","5e6142b297c182e901000217","5e6143b997c182e90100024b","5e6144cb97c182e90100027c","5e614762956b37df01000025","5e6147cb956b37df01000034","5e6147cf956b37df01000036","5e614823956b37df01000047","5e614848956b37df0100004f","5e6148f6956b37df0100006c","5e6149b8956b37df01000081","5e6171e6417d19de0100013a","5e6172fd417d19de01000162","5e617359417d19de01000169","5e61743a417d19de01000184","5e617465417d19de01000189","5e617609417d19de010001aa","5e617696417d19de010001b3","5e6177a5417d19de010001ca","5e6177c2417d19de010001ce","5e617852417d19de010001f0","5e6193c81d4ccede0100010b","5e61a5fa2ef239de01000069","5e61c964abaeaede0100010d","5e6264ec11ac5fde01000066","5e62697811ac5fde01000100","5e62a99408ebcade01000152","5e62ab4d08ebcade0100016e","5e62eb133ba99bdf01000131","5e62eb9d377bb9de01000002","5e633e70716092de010000c8","5e63c4ac74499adf01000108","5e63d62f78da4fdf0100006a","5e6527d9ee6356df01000070","5e659b015dd5c8de01000086","5e676de810be53de0100000a","5e67759110be53de01000072","5e67b2c184cf9ede0100041c","5e67e8d00e29d3de01000164","5e67ef260e29d3de01000223","5e6920fa662319de01000226","5e6971b1ae9c71de0100033c","5e69a3dc23ebccde01000062","5e6b9b4138517edf0100000b","5kT7ojhuB22wGn6pt","5tqh95krD4dKFNZAQ","5x3AEdxLnChwvZa8g","69vAtG6YtnrsSoazH","6EWojj4LmiAwaL6B9","6Ffmir77DS8x3E6sx","6JEETkGR6ZRiAfczM","6TLgrWa44XWGyNXgP","6e2XFy229hLbFR99A","6g5xvCGSh2vtto77h","6huhEvSXfaqHXmkc3","75bFNvjwbxXRG4yCg","7DBAebkzkzghuoYzR","7LxvPnLQ5LmtN2oih","7PnbdWfd9vxcGEmqs","7RXMsKBr98X9uDG4u","7ciJ76kYqGFFFHg2A","7nrMwnG4s5qdFJFZo","7qwQRtBSKxFLoshsT","7sD9QYn38k6CWpDP4","85yCL2SYx2tjZo89P","8eeAa2j6MpLZi53zn","8gia9Wcidjqp4tov5","94Hpgyrdi9jy4XNfi","9JYRiKpufhJJzKFwW","9vADaRieFWKoQPYGA","9x6sTTNqAw2FaeeWG","9yWXZ5ebx92nLDhL2","A7MG4bLJGjd3nScQN","ANKFWzBqWy9E3pEaZ","BSch39FbG778QEiy5","BaD4cmvBYjC63mtCM","BkX37EvGx8eMaKWi2","BtBojnWebnx7wJENc","Bv3Zb4Heu6oao7LnR","C5WLgjYCWGp8LBkmp","CCHfnDadCwL3KoB9E","CLPkk3fjJiHBM3Y8f","ChiohN3JKHSyXQA3m","CiC5mkb4wNBrYR4We","CqY7Kc6gRRDohbtu9","D8TuWoWXbPtuvvx2B","DDevC6ZgEpNgZWH3X","DSTBpPjpYJ7CrMgaG","DY9xcsoTYxu4PcDug","DzywbfM9RtjEziNni","E3FPR2oSan5tuHkyw","E7NDYosNnpGkNTzcR","EEpw2vd5BnZwztHEi","EJHss79aCiBRDZzPu","EKbcyTPbo56hpER8z","ELh3gPsJzYdwvJByj","EcbrbYRcyh2MnmkHB","Edfq2hw3azcDSTkEs","EtCE2kkfYCoWt2Ejw","EvNdDTjjqKFS2AnSi","F26qR6WH5QodtsYb4","G2XSrL9u7Zo4p75tM","G4TxepEm9bNPjAeMw","GQFnFjkwarniwEv3W","Gres6ZKfGnprpeamY","HF68i5BjL9KiE5P7o","HJLuNjpriEJ9QwHXT","HMLjh7735Nf6KaJwA","HcwYCxEWMuy337yxY","HfBtxat7wEXtvmNx4","HgNFZX9da6d5znCcG","J7d7XjpSZNthLLmzm","JK4SPS2cSvf3cPQDP","Js3p25Ebif4kCDGX9","KBQ4a4vXbMzrtd9Ri","KCHLdJic3Xti28ycM","KXgQR7wLS6mAhktez","KbdpDdmpcBK8frGCg","Kibc5hjLCbcH33GZR","Kkh3yAifcZRRazsTE","KtCvweD8r92ANqQ9s","Kv89ueyzCCxX323kg","LSr8rdNmg6JzNc6s3","Lny2nNi3DiKEBXHpH","Lr9vLkup6tgRqb2H6","M9H5ZjDn8qsDKy2HS","MpjduyZMdMMeTCQar","MqntDHkL73WMSYYEf","MtTEKC4XxAQntH2m9","MzrNu3MqSeatEHDuB","N2AiqindFhfTWwzG9","N9tCGBsMcBkr7h3Jf","NFGnmR7f5ZLhR3Y83","NjsqzFAPLtjHWDutm","NuNFTPoNEJsvXiAdw","P2boEvDD9nEtLDcsT","P66zMu2uevumeMyNQ","PEv4bHqvu98SoC7Nr","PGr3aYEWEcRsJTTMM","PZHqjKS2BtKjzq28j","Pad7dGf7zRDQKnfPu","PbKBaE9nWH4yrgQsn","PfxG4GGaqck8CYBgW","PgdLqFJEDcZXck7XY","PyejnXwfbytABY8Ss","QPkFZFs4JtdTMMYeY","QY9z9nWeE7HsheEvM","QcA2w8WB8pskaj36d","QcgXmZjQQehLf48uz","QyCdaaevKJn7zAp7c","R2tXCYgFme4gFKCNC","R9CzgnpThCKSYQYQJ","RBTe6hubxveyMK5HH","RN9Q569t45QtXMRus","RPDNs66wjvQvGCnmW","RX4f9LrmJvdsJgGNq","RutBMyXGKZTCxFmnu","SCDCkFXictWDuwRnP","SDQcntjjZvnPFcWuS","SZjZ9JusAKK99898w","ScZRPZu5ALEREpZNG","SvAeHjtw8uYu3ihPg","TGjnzins6tGHL8EPq","TRnXA3ibpxC4vKmBA","TeRWMwdR8PeyKucsZ","Tfnid3jXw4X5eEAwb","TkMSfrtqiEYcarybr","TpN47khzw8cbrcnkG","W6sSkuAH64Ea4MbEx","W76dvrF39QbxPziTy","W7GKJMGm3FPMm8wN3","W8ge27yKHbWdSnoWv","W93qCNzi5rc6NmKjR","WCiQS7jgbNPDayDoD","WEbj6a2uMBxmTmCZd","WK33EBjTc5HFEx9LY","WdiiwdJgcavsy9dsH","Wza44JooxKS6dBCSY","XE66B9odchJ462xqj","XJAj2ncSnpwZq3Zuv","XQj7AQR5JurFtAY2c","Xmnqb2bDD5TM3DSqK","YLFYg6mMNR3TckeMp","YhFgPXemga3EjrY4C","YkRZrYsisZjGFqn4i","YsG9JDiG3qRSkezZB","YyaFKTK3H8XN77jTo","ZN4Coq5sLaTaqSGkH","ZqkHZRiLR5QK4FFqt","ZrAgNouGZisEYvQ6R","ZuWy2sc6J7Kg5rTaw","a4vfY6tk23HvxuXih","a5mDhx9G7o5DfzwK7","aBiAKt4v4Z8ZNBxHy","aBumSK6KzzzLcQCcD","aQb87jtkZzTpTXmRq","aWmEXD5Bd2HATh5f9","aa43RppmjWcPSNEWn","arer9KQmLSZJyKEjZ","asDCqgPCgWCgd7T5m","azzu48Y9BFosGy6pT","b3kXMriN8W2fvhzGP","b4YuxKEe28fkyofPr","bMpKb8Wx4yXRymKut","bgWkatnia6jr5xayi","bnEjdLAA8CEAzfpS8","borKKcT3xxBpudQ64","bsgRA4rZFfBc8QYoc","cCeCRM6h4ucLQPJ4d","cFdSiDByk4WjY8CAq","cWt6SmxTGrKfAbNHE","cbcqvWgdKjbNqs6mB","ccbF82t7WawnKZ4Rn","cdFMvAfHGmZgC2mf5","co6uTMnYgX9vm8yKt","ctW3gcjN4LJfcbvWL","d4MzZqNqd9RaxYvKg","d8wXTv3AWGbjzpBHg","dH287JHPXSQxmbkaY","dXWXguojmFwXfC343","dcS3nv4fRYtjWdi2q","dmuH2prrwszZRG4Gy","dtxxgovfEi68EZcf8","e4dcY3NNfY2s62sc4","eNrcvXWLvGDMNMTZk","eRagYdajXLmC7qAde","erKe6PYERQP4JjJNS","evFHRXtQiNy2Rbex6","exKqG5EDehFsMCfbj","fGznKFztBFTWKbXcC","fT5agwbbQsrQvHwjr","fbEH2noEwnSJ97K6c","fhgiAXrjjqqpY7MEC","fwa4FWk6o8ifpehjT","fxAe3pLLGAQzMiXtf","g6GQMLLPCYtSEPHJr","gWuDChFbHMsawLBYs","h7nTPA8M5zbg5J7xG","hFXd3NHqjqjt6EYoc","hPyxgq8atsySd5xn2","hSMkzeYwKbFwkzPNf","hbeDvNmTzh9xWZKfE","hd3fRad58ComrShen","hfdKshgYhSgxvHWPG","hnRkpQqfbxyJ3tkn9","hskMGaWwn79efdrq5","iQrP3yuSpq3u5DeGB","iR3L6GjNhtwHoEoFy","iRYtDaK9r3m8jN2Sn","iRiXZZjhB3zha5oX9","iZMPdSMJGv3pWzNuA","iiCzWFNEcEPqwcKvw","iqMpMab3yd9vLzPYB","isSuLydDnaYwiJLKz","j3kRLg9zrogakxiSA","j5XZBLr6PrXYev348","jAL37CjxprJspwLaH","jNvX6sRhoA2H8KcBL","jSeZSdTc7fKqYKPus","jsDXue2QRhPNtXGQ4","kRND6owHip8yssCor","knf5uyxK2ow3SpqCu","kqHwE9tgRXMXu4kEQ","kyq6KHQpzZAdmThEk","mQFywaq3wPjebF2eL","mXh2EpqubDNEMwEjG","mrnsuReH9ustitmYb","n7WdrYhwgBQCu5kr8","nT3poDttXANjYSX6e","nqnh7MctWtHu9Eeri","o5tJCNRb3b7NvZM5q","oLh7xWSFokQyudb2P","ohTwWijM8a9gyzeRF","ourp9XrgdyXj8WFuK","ovSyoqQ7aH52YioMn","pQxuwh4qoozdvFcAc","pd36XDsh7HiCEHrnM","pgXZ828cLEnY2sEGB","piwrxxTYoZJD9vE4v","pjdBNzfX6Fg3iNg3R","pxrWFLGpEpQ42fuBE","q2LtMHAiQoMT8Xchg","qHumEvKrAo59WDsyy","qQ3AasiFibou32YB3","qQGmWXiPhpFtukHky","qWaJKLfpRcauovktX","qZzTbPCMTYnccLMQF","qcv8BfuqP46w5TCgm","qid5XCfkD2WwgLS8R","qyy8E73fj4a6arNav","r7PQ9yjAary28C7nZ","rCk6qGYaiRAdFyk9D","rM5SKh2zyNfWuFtQq","rPbeYkxxnMztRpYJN","rRHCJ6u8mg2RBjEse","rTokX5ZPqXH7AW8Q3","rjMjv9Zs6uN62LBCw","rosk3uCrKbm4zsN9c","rxtH2xFnYvncfWt2t","sPXiTChLNMv3hpXgb","sW5GEKDWvnTewQkbT","sXS57Fc7fgruknxpf","snzsZS4ZoQsssH7jo","t59nTA2vifvAL62fu","tAiJyC4sSeJovKo6g","tEzfQHo2dSBAcxkST","tGstYCAK9uNSeBdMr","tPh5GGfDKphEBoZQh","tQf7bZbHEJ7X2fezF","tf3aDwmjr4hxPxPhj","tgmv2kuNNpCmKkpdZ","uEp67jLjsQBiaC534","uZtecZgDRcNbzjRrS","uhmJ5WhLYgMYerTHi","uiHckiLwyEGaZnMjy","ukjxjyKEfh7uZAsio","uyux8pktYF6ssDoZf","uzkRoqSWcQ6n9Yeyz","v7a3d8AxWiyo9n8Fn","vD49jJLRfcCysRAwG","vGwk5eaop9yEtCJa8","vZw5cZxtwJMvfc58H","vj7jDESLnfMknrEuB","vsMXhpr8ETj9TD5z9","wBP9jvakyv829nNA4","wNDfgWHjeKnqBS9nv","ws6jv4CTkMahvFLhu","xNSyGLEfPK8b9jsnn","xRuBz2yYCKFkHHLXt","xkufTn8vsJ3BtChuK","xmw7Z7X9BxL2T4TXF","yQFLKvbmphB4J32Ju","yY2QrRhTF3HZGJMSj","yY6GA8DJonrxgDYB7","yePcFQx6YsddjgtnJ","yjrJ3sgEPED7nwj9d","yjvi3LGSByST2m4p6","ynu4jgHH2ccNtszqw","zRWju2ET6AQ8fCBJD","zSQERHPnB6FgcG7m9","zWBeaXAxKdgwnXjqh","zX69BSzns4M2sgdcS","zsfGvcy74bAeDKAiM","zwzowH3TqD2cTSYis"],"dataSources":["KjwPBikSY3GLxsjaT","PGvySWpKSvwM9TdfF","dkiwe9ywwzdCpReAj","j35RzudoCYaQbDv7b","PcAmxdFwRBJmeNAnx"]}