MATLAB: Which architecture and settings are the bets to predict the next value of a time series by time delay neural network

neural network

I have this moving window input:
input_normalised =[0.7764,0.7779,0.7900,0.7848,0.7919,0.8071,0.8091,0.7914,0.7794,0.7878,0.7927,0.8039,0.7482,0.7241,0.7950,0.8695,0.8360,0.7766,0.7756,0.7737,0.7881,0.7937,0.7935,0.8065,0.8132,0.7906,0.8514,0.8766,0.8652,0.8093,0.8228,0.8186,0.8137,0.8225,0.8072,0.8034,0.7904,0.7734,0.7718,0.7852,0.8051,0.8164,0.8305,0.8142,0.7913,0.7902,0.7873,0.7869,0.7808,0.7932,0.8096,0.8256,0.8125,0.7791,0.7717,0.7807,0.7906,0.8244,0.8379,0.8224,0.7876,0.7953,0.8126,0.8202,0.8127,0.7957,0.7987,0.8173,0.8174,0.7910,0.7770,0.7851,0.7994,0.7998,0.7839,0.8970,0.9239,0.8819,0.7718,0.7945,0.8205,0.8190,0.8567,0.8489,0.8315,0.7791,0.8218,0.6526,0.6428,0.7157,0.8779,0.7963,0.8042,0.7888,0.7855,0.7778,0.7901,0.7857,0.7903,0.8064,0.7521,0.7338,0.7637,0.7821,0.7302,0.7374,0.8007,0.8071,0.8467,0.8477,0.8400,0.7931,0.7920,0.7893,0.7847,0.7834,0.7759,0.7957,0.7894,0.8490,0.8535,0.8695,0.8253,0.8392,0.8171,0.7856,0.7607,0.7803,0.8198,0.8009,0.8013,0.8209,0.7487,0.7194,0.7452,0.8143,0.8778,0.4527,0.3091,0.4690,0.8998,0.8130,0.7980,0.8041,0.7955,0.7843,0.8472,0.8771,0.8267,0.7525,0.7778,0.7977,0.7720,0.7684,0.7767,0.8577,0.8901,0.8723,0.8022,0.7995,0.7887,0.7873,0.7880,0.7982,0.7741,0.7747,0.7888,0.7887,0.7972,0.8055,0.8168,0.7780,0.8036,0.8125,0.8114,0.8034,0.8053,0.8015,0.7964,0.8088,0.7846,0.7812,0.7644,0.7701,0.7650,0.7898,0.8067,0.8129,0.7974,0.7794,0.7905,0.8122,0.8217,0.7851,0.9257,0.9855,0.9309,0.7744,0.7885,0.7817,0.8170,0.8406,0.8184,0.8229,0.8474,0.8300,0.8285,0.8430,0.8350,0.7920,0.7849,0.8465,0.8976,0.8853,0.8177,0.8105,0.8086,0.7945,0.7801,0.7849,0.8044,0.8143,0.8229,0.8228,0.8060,0.7923,0.7905,0.8411,0.8555,0.8270,0.7810,0.7597,0.9403,1.0000,0.9472,0.7639,0.7867,0.8291,0.6682,0.5983,0.6473,0.8202,0.7888,0.7863,0.7860,0.7943,0.7846,0.7815,0.7956,0.8042,0.8286,0.8308,0.8103,0.7866,0.7952,0.8023,0.7888,0.7935,0.8132,0.7854,0.7665,0.7747,0.8045,0.7737,0.8739,0.9140,0.8713,0.7874,0.8256,0.7859,0.7508,0.7704,0.8072,0.7870,0.7765,0.7871,0.7973,0.7904,0.7965,0.7904,0.7844,0.7813,0.7980,0.7888,0.7661,0.7646,0.7964,0.7776,0.7589,0.7722,0.7851,0.8010,0.8160,0.8127,0.8640,0.8963,0.8642,0.9245,0.2348, 0,0.2497]
and I want to get this value by prediction of the neural network:
real_normalised_next_value = 0.9643

Best Answer

If you only have 1 series, use NARNET
help NARNET
doc NARNET
Search BOTH the NEWSGROUP and ANSWERS using
greg NARNET
Hope this helps.
Thank you for formally accepting my answer
Greg